EDUCATION IN EMERGENCIES. EDITORIAL NOTE Editorial Board

Transcription

1 VOLUME 3 NUMBER 1 JULY 2017 JOURNAL ON EDUCATION IN EMERGENCIES EDITORIAL NOTE Editorial Board EiE RESEARCH ARTICLES The 4Rs Framework: Analyzing Education s Contribution to Sustainable Peacebuilding with Social Justice in Conflict-Affected Contexts Mario Novelli, Mieke T. A. Lopes Cardozo, and Alan Smith Can Teacher-Training Programs Influence Gender Norms? Mixed-Methods Experimental Evidence from Northern Uganda Marjorie Chinen, Andrea Coombes, Thomas De Hoop, Rosa Castro-Zarzur, and Mohammed Elmeski The Limits of Redistributive School Finance Policy in South Africa Rachel Hatch, Elizabeth Buckner, and Carina Omoeva EiE FIELD NOTE The Potential of Conflict-Sensitive Education Approaches in Fragile Countries: The Case of Curriculum Framework Reform and Youth Civic Participation in Somalia Marleen Renders and Neven Knezevic BOOK REVIEWS Childhood Deployed: Remaking Child Soldiers in Sierra Leone by Susan Shepler Jesper Bjarnesen Training for Model Citizenship: Ethnography of Civic Education and State- Making in Rwanda by Molly Sundberg S. Garnett Russell Partnership Paradox: The Post-Conflict Reconstruction of Liberia s Education System edited by Christopher Talbot and Aleesha Taylor Laura Quaynor Critical Peace Education and Global Citizenship by Rita Verma Lynn Davies July

2 JOURNAL ON EDUCATION IN EMERGENCIES Editor-in-Chief: Dana Burde, PhD, New York University Managing Editor: Heddy Lahmann, New York University Book Review Editor: Elisabeth King, PhD, New York University Editorial Board: Carine Allaf, PhD, Qatar Foundation International Ragnhild Dybdahl, PhD, Oslo and Akershus University College of Applied Sciences Mark Ginsburg, PhD, University of Maryland Elisabeth King, PhD, New York University Mary Mendenhall, EdD, Teachers College, Columbia University Susan Garnett Russell, PhD, Teachers College, Columbia University Sweta Shah, PhD, Bernard Van Leer Foundation James Williams, EdD, George Washington University Inter-Agency Network for Education in Emergencies c/o International Rescue Committee 122 East 42nd Street, 14th floor New York, NY 10168, USA New York University International Education Program 246 Greene Street, Suite 300 New York, NY Editorial Assistants: Laura Lopez-Blazquez, New York University Pamela Montalbano, New York University Sarah Haroon Sualehi, New York University Copy Editor: Dody Riggs Editorial Designer: Erik Ramirez ISSN Open-source publication; distribution free of charge; available in soft copy only. Copyright 2017, Inter-Agency Network for Education in Emergencies. All rights reserved.

3 JOURNAL ON EDUCATION IN EMERGENCIES VOLUME 3, NUMBER 1 JULY 2017 EDITORIAL NOTE Editorial Board...5 EiE RESEARCH ARTICLES The 4Rs Framework: Analyzing Education s Contribution to Sustainable Peacebuilding with Social Justice in Conflict-Affected Contexts...14 Mario Novelli, Mieke T. A. Lopes Cardozo, and Alan Smith Can Teacher-Training Programs Influence Gender Norms? Mixed-Methods Experimental Evidence from Northern Uganda Marjorie Chinen, Andrea Coombes, Thomas De Hoop, Rosa Castro-Zarzur, and Mohammed Elmeski The Limits of Redistributive School Finance Policy in South Africa Rachel Hatch, Elizabeth Buckner, and Carina Omoeva EiE FIELD NOTE The Potential of Conflict-Sensitive Education Approaches in Fragile Countries: The Case of Curriculum Framework Reform and Youth Civic Participation in Somalia Marleen Renders and Neven Knezevic BOOK REVIEWS Childhood Deployed: Remaking Child Soldiers in Sierra Leone by Susan Shepler Jesper Bjarnesen Training for Model Citizenship: Ethnography of Civic Education and State-Making in Rwanda by Molly Sundberg S. Garnett Russell Partnership Paradox: The Post-Conflict Reconstruction of Liberia s Education System edited by Christopher Talbot and Aleesha Taylor Laura Quaynor Critical Peace Education and Global Citizenship by Rita Verma Lynn Davies CONTRIBUTORS July

4

5 EDITORIAL NOTE Editorial Board Every theme that appears in the Journal on Education in Emergencies is by definition timely and important. In this special issue of JEiE, we focus on education and peacebuilding. Although humanitarians for many years credited education with creating or fostering peace, the understanding of this relationship grew more nuanced at the start of the 2000s, with the publication of The Two Faces of Education in Ethnic Conflict (Bush and Saltarelli 2000). This publication ushered in a period of reflection on the role of education, not just in promoting peace but in promoting conditions for conflict through uneven access, biased classroom practices, or negative ideological content (Brock 2011; Burde 2014; Burde, Kapit, Wahl, Guven, and Skarpeteig 2017; Gross and Davies 2015; King 2014; Ostby and Urdal 2010; Shields and Paulson 2015; Smith 2007). Although educators were as committed as always to the transformative potential and power of education, they were no longer as sanguine about its delivery or content in humanitarian crises. At the same time, however, educators continued to pursue a deeper understanding of how, and under what conditions, education could promote peace or contribute to peacebuilding. Increased reflection among practitioners helped refocus this line of research on a more granular and systematic understanding of the mechanisms that could increase underlying conditions for peace. How might gender norms, for example, contribute to or undermine efforts to promote conditions considered important for peace? What are the implications of redistributive school financing for social cohesion? With this new issue of JEiE which consists of three research articles, one field note, and four book reviews we return to the positive face of education as we examine its contributions to peacebuilding. The articles in this issue bring a range of analyses to this question, including a focus on social justice, reconciliation, inclusion, gender norms, and the importance of social cohesion. We present a short description of each piece below. In The 4Rs Framework: Analyzing Education s Contribution to Sustainable Peacebuilding with Social Justice in Conflict-Affected Contexts, authors Mario Novelli, Mieke T. A. Lopes Cardozo, and Alan Smith present an analytical framework for investigating and understanding peacebuilding through education. The 4Rs framework emerged during the early stages of the UNICEF s Peacebuilding, July

6 EDITORIAL BOARD Education, and Advocacy program (PBEA also known as Learning for Peace), a four-year initiative established in 2012 and funded by the Government of the Netherlands. The framework played an important role in shaping thinking on the relationship between education and peacebuilding within UNICEF and in PBEA s evolution. In the article, the authors identify how the 4Rs framework combines dimensions of recognition, redistribution, representation, and reconciliation... to demonstrate what sustainable peacebuilding might look like in postconflict environments (18). They argue that the roots of conflict must be addressed in order for a society to transition to sustainable peace, and that education can play an important role in this process, particularly through social justice and reconciliation. They then apply the 4Rs to a case study in Myanmar, illustrating how the framework provides a useful heuristic device for analyzing peacebuilding and education, which may be applied to academic research, policy-making, and program design and evaluation. In Can Teacher-Training Programs Influence Gender Norms? Mixed-Methods Experimental Evidence from Northern Uganda, Marjorie Chinen, Andrea Coombes, Thomas De Hoop, Rosa Castro-Zarzur, and Mohammed Elmeski detail a mixed-methods study of a teacher-training program on gender socialization in Karamoja, Uganda. The authors emphasize the role education plays in promoting gender equality in conflict-affected environments, and the implications for peacebuilding. Quantitative evidence demonstrates that the program had positive effects on teachers knowledge and attitudes but did not affect their practices in the short term. There was no quantitative evidence for the effectiveness of a complementary text-message intervention. Qualitative data suggests that teachers still identified with traditional gender norms and beliefs about gender (71) and that, although teachers engaged in basic program practices taught during the training, they did not engage in more complex practices. Their findings indicate that, while teacher training can influence knowledge and attitudes toward gender equality, traditional gender norms can be a barrier to changing behavior in the short term (46). The Limits of Redistributive School Finance Policy in South Africa, by Rachel Hatch, Elizabeth Buckner, and Carina Omoeva, focuses on the effectiveness and perceived effectiveness of South Africa s no-fee school policy. This mixed-methods study draws on household and school survey data and qualitative interviews to examine if and how the no-fee school policy has contributed to equity. Their findings show that the policy has reduced the financial burden on many black households, which are often in poorer communities (80), but that gains have not been equalizing, and gaps in resources remain (100). Thus they argue that 6 Journal on Education in Emergencies

7 EDITORIAL NOTE South Africa s current school finance policies may be better characterized as pro-poor than redistributive, and point to implications for social cohesion (79). In our one field note for this issue, The Potential of Conflict-Sensitive Education Approaches in Fragile Countries: The Case of Curriculum Framework and Youth Civic Participation in Somalia, Marleen Renders and Neven Knezevic describe an innovative education intervention aimed at addressing the drivers of conflict in Somalia. The program directly engaged local youth in a participatory process to contribute to a review of the primary school curriculum framework. This approach involving youth in the facilitation of local communities discussions of curriculum changes has the potential to address historic legacies of authoritarian national governments and top-down governance systems and to offer many traditionally excluded groups an opportunity to contribute to a national development process (125). The authors contend that education interventions that directly engage with the factors driving conflict and that directly involve the voices of a wide range of stakeholders, particularly local communities, have important implications for peacebuilding. We also include four book reviews in this issue. In the first, Jesper Bjarnesen discusses Childhood Deployed: Remaking Child Soldiers in Sierra Leone by Susan Shepler, published by New York University Press. Shepler examines the challenge of reintegrating child soldiers and youth combatants into society, based on her extensive ethnographic research in Sierra Leone. Bjarnesen notes that the book is particularly relevant for anyone interested in understanding the nuance and complexity of the interface between international conventions on the rights of the child and local notions of childhood and youth in a place like Sierra Leone (130). In the second review, S. Garnett Russell examines Molly Sundberg s Training for Model Citizenship: Ethnography of Civic Education and State-Making in Rwanda, published by Palgrave Macmillan. Drawing on Sundberg s ethnographic research, combined with her experience as a development practitioner, the book investigates the role of Rwanda s state-sponsored civic education program. Russell notes that the book offers an in-depth portrait of an important topic in postgenocide Rwanda and sheds light on the subtle contradictions and tensions of active citizenship in a postconflict authoritarian state (134). The third book, reviewed by Laura Quaynor, is Partnership Paradox: The Post-Conflict Reconstruction of Liberia s Education System, edited by Christopher Talbot and Aleesha Taylor, published by Open Society Foundations. This open-source publication provides a critical look at postconflict aid partnerships in Liberia between 2007 and Quaynor writes that many of the contributing authors offer public critique of themselves, their organizations, and others; their willingness to share insider information on July

8 EDITORIAL BOARD the formation and navigation of such partnerships can best be described as brave (136). She considers this volume a must-read for practitioners, academics, and students of EiE. Our last reviewer, Lynn Davies, discusses Critical Peace Education and Global Citizenship by Rita Verma, published by Routledge. The book discusses ways educators incorporate peace education into classrooms through unofficial curricula, and the importance of stimulating peace activism. Davies notes that this book graphically shows us that the key task for our time is not learning about peace but learning not to hate (142). The idea for this issue of JEiE originated with UNICEF s PBEA program, and a number of the articles that appear emerged from projects that were funded by this initiative. In recognition of UNICEF s role in this work, we offer a short description of the PBEA program below. UNICEF PBEA 1 Conflict affects children and families directly (by causing disability, displacement, and death), and indirectly (by creating instability, loss of livelihood, and destruction of assets). Both direct and structural violence pose a risk to the development potential of young children (Punamäki 2014; see also Dawes and van der Merwe 2014). According to the World Bank, a child in a fragile or conflict-affected state is nearly three times as likely to be out of primary school, twice as likely to be undernourished, and nearly twice as likely to die before their fifth birthday as a child in another developing country (World Bank 2011). In 2014 alone, the lives of 15 million children were disrupted by conflict in the Central African Republic, Iraq, South Sudan, the State of Palestine, Syria, and Ukraine (Gladstone 2014). The United Nations Children s Fund (UNICEF) mandate includes protecting the rights of children in both emergencies and humanitarian contexts, both natural and human-made. Faced with increasingly complex humanitarian crises that place children and women at significant risk, UNICEF responds to more than two hundred emergencies every year (UNICEF 2013). UNICEF predicts that the caseload for humanitarian action will continue to grow in the upcoming decades. Close to two-thirds of the world s poor will be living in fragile states by 2030, when the UN Sustainable Development Goals come to an end; the majority of them will be young people (Office of Economic Cooperation and Development [OECD] 2015). The combination of climate change, natural 1 Thanks to Friedrich Affolter for contributing this description of UNICEF s PBEA work. 8 Journal on Education in Emergencies

9 EDITORIAL NOTE disasters, conflicts, chronic poverty, and weak institutions are creating new risks and exacerbating existing ones, and over time they may erode peace, reverse development gains, and create new humanitarian needs (World Economic Forum 2016). International aid agencies need to deliver context-specific programs in complex environments that are fit-for-purpose and focus on prevention, including conflict prevention (OECD 2011). In 2011, UNICEF s Evaluation Office commissioned a study to examine the role of education in peacebuilding in postconflict settings. Given concern about frequent relapses into conflict, the study inquired whether and how education interventions and programming could play a stronger role in the peacebuilding architecture of the UN system (Novelli and Smith 2011, 3). One conclusion of the study was that UNICEF should move away from generic programming and toward education interventions that are informed by high-quality conflict analysis and sensitive to local contexts, while also leveraging the education sector s transformative potential in postconflict societies (Novelli and Smith 2011, 37). From 2012 to 2016, with generous funding from the Government of the Netherlands, UNICEF implemented PBEA in order to experiment with and demonstrate whether and how education, as a social service, can help to strengthen resilience, social cohesion, and human security in conflict-affected contexts, including countries at risk of, experiencing, and recovering from conflict. The countries participating were Burundi, Chad, Côte d Ivoire, Democratic Republic of Congo, Ethiopia, Liberia, Myanmar, Pakistan, Sierra Leone, Somalia, South Sudan, the State of Palestine, Uganda, and Yemen (Shah 2016). An independent program evaluation carried out in fall 2015 concluded that PBEA s choice of using a social service such as education to deliver peacebuilding results was the right one, and that its emphasis on programming based on conflict analysis led to responsive, context-specific programs that contribute to peacebuilding (UNICEF 2015, 14). One distinctive feature of PBEA was its mandate to collect evidence that would illustrate how social service providers and education in particular can mitigate drivers of conflict. 2 This special issue of the Journal of Education in Emergencies presents research conducted as part of the PBEA mandate. These articles underscore the importance of agencies efforts to deliver services in a manner that is sensitive to conflict and, where possible, in ways that not only effectively address human and socioeconomic development needs but simultaneously transform interpersonal and intergroup relationships (McCandless 2012). 2 PBEA research products can be accessed online at July

10 EDITORIAL BOARD This special issue of JEiE is funded by Education Above All/Protecting Education In Crisis. We are grateful for their support. REFERENCES Brock, Colin Education and Conflict: A Fundamental Relationship. In Education, Conflict and Development, edited by Julia Paulson. Oxford, UK: Symposium Books Ltd., Burde, Dana Schools for Conflict or for Peace in Afghanistan. New York: Columbia University Press. Burde, Dana, Amy Kapit, Rachel L. Wahl, Ozen Guven, and Margot Igland Skarpeteig Education in Emergencies: A Review of Theory and Research. Review of Educational Research 87 (3): Bush, Kenneth, and Diana Saltarelli, eds The Two Faces of Education in Ethnic Conflict: Towards a Peacebuilding Education for Children. Florence, Italy: UNICEF Innocenti Research Centre. Dawes, Andy, and Amelia van der Merwe Structural Violence and Early Childhood Development. In Pathways to Peace: The Transformative Power of Children and Families (Strüngmann Forum Reports, Vol. 15), edited by James F. Leckman, Catherine Panter-Brick, and Rima Salah. Cambridge, MA: MIT Press. Gladstone, Rick. 2014, December 8. UNICEF Calls 2014 One of Worst Years for Children. New York Times. unicef-calls-2014-one-of-worst-years-on-record-for-worlds-children.html. Gross, Zehavit, and Lynn Davies, eds The Contested Role of Education in Conflict and Fragility. New York: Springer. King, Elisabeth From Classrooms to Conflict in Rwanda. New York: Cambridge University Press. 10 Journal on Education in Emergencies

11 EDITORIAL NOTE McCandless, Erin Peace Dividends and Beyond: Contributions of Administrative and Social Services to Peacebuilding. New York: United Nations Peacebuilding Support Office. peace_dividends.pdf. Novelli, Mario, and Alan Smith The Role of Education in Peacebuilding: A Synthesis Report of Findings from Lebanon, Nepal and Sierra Leone. New York: UNICEF. PeacebuildingSynthesisReport.pdf. Office of Economic Cooperation and Development (OECD) Conflict and Fragility: International Engagement in Fragile States. Can t We Do Better? Paris: OECD Publishing. Office of Economic Cooperation and Development (OECD) States of Fragility 2015: Meeting Post-2015 Ambitions. Paris: OECD Publishing. Østby, Gudrun, and Henrik Urdal Education and Civil Conflict: A Review of the Quantitative, Empirical Literature (Background paper prepared for the Education For All Global Monitoring Report 2011: The Hidden Crisis: Armed Conflict and Education ). Paris: UNESCO. Punamäki, Raija-Leena Mental Health and Development among Children Living in Violent Conditions: Underlying Mechanisms for Promoting Peace. In Pathways to Peace: The Transformative Power of Children and Families (Strüngmann Forum Reports, Vol. 15), edited by James F. Leckman, Catherine Panter-Brick, and Rima Salah. Cambridge, MA: MIT Press, Shah, Ritesh Peacebuilding, Education and Advocacy in Conflict-Affected Contexts Programme (PBEA): UNICEF Programme Report Shields, Robin, and Julia Paulson Development in Reverse? A Longitudinal Analysis of Armed Conflict, Fragility and School Enrollment. Comparative Education 51 (2): Smith, Alan Education and Conflict: An Emerging Field of Study. In Education, Conflict and Reconciliation: International Perspectives, edited by Fiona Leach and Máiréad Dunne. Frankfurt: Peter Lang, July

12 EDITORIAL BOARD UNICEF UNICEF Evaluation of UNICEF s Peacebuilding, Education and Advocacy Programme (PBEA). New York: United Nations Children Fund, eccnetwork.net/wp-content/uploads/final-learning-for-peace-evaluation- Summary.pdf. World Bank World Development Report 2011: Conflict, Security and Development. Washington, DC: International Bank for Reconstruction and Development, 5. WDR2011_Full_Text.pdf World Economic Forum Global Risks Report 2016 (11 th ed.). Geneva: World Economic Forum. 12 Journal on Education in Emergencies

13 NOTE FROM EDUCATION ABOVE ALL EAA The Education Above All Foundation, through its advocacy program Protecting Education in Insecurity and Conflict (PEIC), is pleased to partner with the NYU editorial team and the Inter-Agency Network for Education in Emergencies to produce the Journal of Education in Emergencies. Education Above All, through PEIC, is funding the publication of two editions of the journal, including the current one. Our partnership will help build communities of practice and provide thought leadership, including research, analysis, and knowledge generation, in the critical and emerging area of education in emergencies. Our partnership aligns with PEIC s core mandate: advocacy in defense of the right to education during war, conflict, and insecurity. July

14 THE 4RS FRAMEWORK: ANALYZING EDUCATION S CONTRIBUTION TO SUSTAINABLE PEACEBUILDING WITH SOCIAL JUSTICE IN CONFLICT-AFFECTED CONTEXTS Mario Novelli, Mieke T. A. Lopes Cardozo, and Alan Smith 1 ABSTRACT This paper lays out a theoretical and analytical framework for researching and reflecting on the peacebuilding role of education in conflict-affected contexts. The 4Rs framework recognizes that working toward positive peace (Galtung 1976, 1990) requires working toward peace with social justice and reconciliation, challenging dominant security-first and liberal peace models, and gaining a better understanding of how education might support these processes in building sustainable and peaceful postconflict societies. The 4Rs framework combines dimensions of recognition, redistribution, representation, and reconciliation to explore what sustainable peacebuilding might look like through a social justice lens. The paper addresses the cultural translation of these concepts, highlighting the need for locally embedded interpretations. Rather than a fixed theoretical model, the 4Rs approach is designed as a heuristic device that promotes a dialogue among key stakeholders on the dilemmas and challenges in the field of education in emergencies. We highlight the application of a 4Rs framework through a recent case study of Myanmar, which demonstrates both the interrelated connections and the tensions between the different Rs. Finally, we reflect on the challenges and limitations of the approach, and the tasks ahead. Received January 22, 2016; revised September 30, 2016, and December 21, 2016; accepted January 30, 2017; electronically published July Journal on Education in Emergencies, Vol. 3, No. 1 Copyright 2017 by the Inter-Agency Network for Education in Emergencies (INEE). All rights reserved. ISSN Corresponding authors: t.a.lopescardozo@uva.nl; m.novelli@sussex.ac.uk; a.smith@ulster.ac.uk 14 Journal on Education in Emergencies

15 THE 4RS FRAMEWORK INTRODUCTION AND RATIONALE OF THE 4RS FRAMEWORK While education is a core demand of communities affected by conflict and a crucial element in recovering from war and building sustainable peace at various levels, it is often seen as a soft measure that can be put aside. As such, it remains of marginal concern to the major United Nations and other international agencies tasked with promoting peacebuilding. Responding to this reasoning, several nongovernmental and governmental actors have campaigned for the contrary view: that education cannot wait, especially not in contexts ravaged by conflict and other types of emergencies. We extend this broader argument that education cannot wait by focusing on the crucial role education plays in promoting sustainable peacebuilding. The overarching rationale for our approach is underpinned by a broad definition of the long-term objective of education and peacebuilding interventions that is, promoting peace with social justice and reconciliation as well as the role education can play therein. Previous research has led us to recognize that working toward positive peace (Galtung 1976, 1990) requires working toward peace with social justice and gaining a better understanding of how education might support these processes in building sustainable and peaceful postconflict societies. It also has made us aware of the complex challenges faced by policy-makers and practitioners who are seeking to expand the role of education in peacebuilding activities. In this article, we build on our previous work on the role of education in peacebuilding (Smith 2005; Novelli and Lopes Cardozo 2008, 2012; Smith, McCandless, Paulson, and Wheaton 2011; Novelli and Lopes Cardozo 2012; Novelli, Valiente, Higgins, and Ugur 2014; Lopes Cardozo and Hoeks 2014; Novelli, Higgins, Ugur, and Valiente 2014; Lopes Cardozo and Shah 2016) by presenting an analytical model that reaches beyond academic analytical relevance. This model tends to be of more practical use in the planning and evaluation phases of policy and programming in social service delivery. This model was specifically developed as part of the work of the Research Consortium on Education and Peacebuilding, which was supported by UNICEF s Peacebuilding, Education, and Advocacy (PBEA) program between July 2014 and June The work was led by the universities of Amsterdam, Sussex, and Ulster and co-directed by the authors of this article. 2 The consortium, which sought 2 The consortium has emerged out of a longstanding relationship between the authors, UNICEF, and the Netherlands Ministry of Foreign Affairs that stretches back to The UNICEF PBEA program is a $200 million, four-year partnership ( ) between UNICEF, the Government of the Netherlands, the national governments of 14 participating countries, and other key supporters. July

16 NOVELLI, LOPES CARDOZO, AND SMITH to gain knowledge on the relationship between education and peacebuilding in conflict-affected contexts, carried out research in Myanmar, Pakistan, South Africa, and Uganda. The research focused on three key thematic areas: (1) the integration of education in UN peacebuilding missions and frameworks, and the integration of peacebuilding in national education systems, policies, and programs; (2) teachers role in peacebuilding in contexts of conflict; and (3) education s role in peacebuilding initiatives that involve youth in contexts of conflict. The research, which was completed in partnership with colleagues in each of the participating countries, aimed to contribute to theory and practice in the field of education and peacebuilding, and to develop theoretically informed, policy-relevant outputs (Research Consortium on Education and Peacebuilding 2014). To lay the foundation for the discussion that follows, we first need to address the problematic nature of the term peacebuilding itself, which has become an increasingly slippery term that is employed by a variety of actors for a wide range of political projects: maintaining security, ensuring stabilization, and other, more transformational processes (Barnett, Kim, O Donnell, and Sitea 2007). This reflects the contested nature of the concept and the historical evolution of debates regarding peacebuilding, particularly as related to Galtung s (1976) notions of negative and positive peace, and the different agendas of actors involved in peacebuilding across the world. 3 For some actors, particularly those working from a humanitarian or security-first approach, peacebuilding denotes a narrow set of activities aimed at ensuring stability in the immediate aftermath of a conflict. For others, peacebuilding represents a more transformational agenda that takes place over a much longer timespan (for a review of approaches, see Heathershaw 2008; Richmond, Björkdahl, and Kappler 2011). Clearly, while acknowledging that actors approaches are situated along a continuum, the role education plays might look very different, depending on various conceptualizations of peacebuilding. These different discursive and often context-specific understandings of peacebuilding are important, as the various actors pursuing disparate interpretations have unequal power and influence. Due to the highly contested nature of peacebuilding, we have found it necessary to develop a normative framework for what we consider the core dimensions of a socially just postconflict society that is heading toward sustainable peace and 3 Galtung (1976) introduced an important distinction between negative peace (the absence of violence) and positive peace (the absence of structural violence and the conditions for war). He distinguishes between three forms of violence. Direct violence refers to physical injury inflicted on another human being. Structural violence is more indirect, is built into the structures of society, and shows up as social injustice and unequal life chances. Cultural violence involves any cultural norms, beliefs, and traditions that make other types of violence seem legitimate, accepted, normal, or natural. 16 Journal on Education in Emergencies

17 THE 4RS FRAMEWORK reconciliation. In our approach to sustainable peacebuilding, we argue for a greater emphasis on social development including education to address the underlying causes of conflict, such as political, economic, and sociocultural inequality and injustice. Political, economic, social, and cultural transformation are needed in conflict-affected societies to support positive peace, and to address rather than reproduce or sustain the injustice and inequality that largely drive conflicts. Transformations are described in terms of the extent to which education policies, individual and institutional agency, and development programs promote the social justice dimensions of redistribution, recognition, and representation, as defined by Nancy Fraser (2005), complimented by postconflict issues of reconciliation (Lederach 1995, 1997; Hamber 2007, 2009). These four elements constitute our 4Rs framework. We contend that, when education applies these multidimensional elements to injustice, it can contribute effectively to what Fraser terms a transformative remedy. Rather than overcoming social injustice with so-called affirmative remedies, which correct outcomes without changing structural frameworks or the status quo, Fraser (1995, 82, 86) argues for transformative remedies that correct outcomes by restructuring the underlying generative framework. We see this transformative emphasis as closely connected to the notion of sustainable peacebuilding. Our basic claim is that education can play an important role in fostering positive peace and social justice, both of which are necessary to transform the root causes of conflict. Hence, our analytical model includes a continuum that ranges from negative peace or the mere absence of violence on one end, to positive peace on the other end. This continuum provides us with a normative scale or lens that we can use to analyze and review education and peacebuilding policies and programs. While normative, our 4Rs model aims to be broad and inclusive, and to recognize that each of these dimensions needs to be translated and embedded in particular local and national geographies, which we further elaborate in later sections. In other words, by developing and applying this 4Rs framework, we claim that the key postconflict transformations needed to produce sustainable peace or, as Galtung (1990) refers to it, positive peace involve redistribution, recognition, and representation. These factors, together with issues of postconflict reconciliation that are linked to transitional justice and dealing with the legacies of conflict, will help bring about greater social justice, as suggested by Fraser (2005). We highlight these four key messages in Textbox 1. July

18 NOVELLI, LOPES CARDOZO, AND SMITH textbox 1: four key messages from the 4rs framework Our theoretical framework contends the following: A sustainable approach to peacebuilding emphasizes social development and addresses the underlying causes of conflict, such as political, economic, and sociocultural inequality and injustice. Education can make a significant contribution to sustainable peacebuilding by providing greater security, as well as political, economic, social, and cultural transformations within conflict-affected societies. Transformation refers to the extent to which education policies, individual and institutional agency, and development programs promote redistribution, recognition, representation, and reconciliation the 4Rs. We need to acknowledge the politics and other complex factors at play in the close interconnections among the 4Rs. We have developed the 4Rs approach as a heuristic device that supports the process of design, data collection, and analysis in order to reflect on the dilemmas and contradictions inherent in supporting the positive role education plays in peacebuilding. Our aim is that this framework will spark a dialogue among key stakeholders and be adapted in ways relevant to each cultural, political, and economic context. This article has a threefold structure. We first critique the dominant security-first and liberal peace peacebuilding models by showing how they fail to support positive peace, and then lay out the potential, and the challenges, for education to play a greater role in peacebuilding processes. Second, we propose an alternative theoretical and analytical model that puts education at the center of building sustainable peace with social justice. We identify how the 4Rs framework combines dimensions of recognition, redistribution, representation, and reconciliation, and examine the work of Nancy Fraser (1995, 2005), Johan Galtung (1976, 1990), and John Paul Lederach (1995, 1997), among others, to demonstrate what sustainable peacebuilding might look like in postconflict environments. And third, we 18 Journal on Education in Emergencies

19 THE 4RS FRAMEWORK illustrate the methodological opportunities and challenges in applying this model to the work of the Research Consortium on Education and Peacebuilding and to a recent case study of Myanmar. This third section aims to operationalize our analytical framework in practical terms by critically analyzing education policy and programs to show the interrelated connections and tensions among the 4Rs. We close by reflecting on the importance of theory-building and development in the field of education in emergencies. EDUCATION AND PEACEBUILDING: FROM A CRITIQUE OF THE FIELD TO A CRITICAL METHODOLOGICAL APPROACH In this section, we analyze the shortcomings of the hegemonic approaches that currently dominate the field of peacebuilding, namely, the security-first and liberal peace theses. We then contrast these with the theoretical literature that informed our rethinking of what sustainable peacebuilding could look like and helped to shape the theoretical and methodological approach that frames our research on education and peacebuilding. As described above, one s approach to peacebuilding depends on one s conceptualization of it, which concurrently informs the role one foresees for social development in these processes, including education. Importantly, there is clear evidence of an imbalance of power between actors operating on different geographical scales. This is reflected in tensions between setting agendas, formulating national policy, and implementing phases of the policy cycle. There is a strong sense of global agendas trumping national priorities, while local needs are marginalized and sidelined. Realities and priorities appear to be highly divergent, and while we can clearly trace global policies that filter downward through the policy cycle, evidence of upward feedback that reflects more bottomup participation is less prevalent (Novelli, Higgins, Ugur, and Valiente 2014). One example of this is the security-first agenda, which is closely linked to the implementation of what Paris (2004, 2010) calls the liberal peace thesis. The liberal peace thesis prioritizes the introduction of liberal democracy and market forces as key drivers of stability, once security has been achieved. According to Castañeda (2009), this can be conceptualized as a trickle-down approach to peace, where the aim is to first achieve a negative peace (cessation of violence) and to then introduce representative democracy. The idea is that these two factors will encourage foreign direct investment and lead to economic growth. However, just as trickle-down economics failed to reach many of the July

20 NOVELLI, LOPES CARDOZO, AND SMITH most vulnerable populations in the 1980s when the International Monetary Fund and World Bank promoted structural adjustment policies, it is now not clear that trickle-down peace is a sufficiently robust development model to reach the most marginalized populations. It may in fact contain the seeds of continuing insecurity (Duffield 1998, 10; see also Paris 2004; Pugh, Cooper, and Turner 2011; Richmond and Mitchell 2012). This global agenda frames much of the international discourse on peacebuilding and, according to Paris (2010), has received wide-ranging critiques over the past decade. While we recognize that the liberal peacebuilding model should not be viewed as a unitary and homogenous model (see Selby 2013), our critical analysis of its core rationales can help us understand why UN peacebuilding programs investment in the social services health, education, and welfare lags behind investment in promoting security and democracy (McCandless 2011). The prioritization of security and the marginalization of education were evidenced in a three-country UNICEF study of the relationship between education and peacebuilding in Sierra Leone, Lebanon, and Nepal (Smith, McCandless, Paulson, and Wheaton 2011; Zakharia 2011; Novelli 2011a, 2011b; Vaux 2011; Novelli and Smith 2011). The findings of this study indicated that the major international actors involved in peacebuilding (UN Peacebuilding Support Office, Department for International Development, USAID, among others) prioritized security, democracy, and free market issues, particularly in the early to medium postconflict phases. They did so at the expense of social-sector spending. The rationale underpinning the prioritization of these issues is that security is the foundation on which development can occur. Denney (2011) notes the following in her research into Department for International Development activities in Sierra Leone: Security first denotes the idea that before one can sustainably engage in development, a basic level of security must be established. A secure environment will ensure that development efforts are less likely to be disrupted or diverted by conflict, and that stability will attract investors who would otherwise be dissuaded by volatility. In this way, security is a precondition of development. (279) Denney (2011) suggests that security and development do not occur symbiotically, that it instead appears increasingly that development has not followed security, which has led to the uncomfortable coexistence of security and what she terms misery. Acknowledging this uneasy coexistence, the research conducted in Sierra 20 Journal on Education in Emergencies

21 THE 4RS FRAMEWORK Leone, Lebanon, and Nepal suggests that, while security in postconflict situations is clearly important, it is not a sufficient condition to reach positive peace or support the social transformation necessary to ensure that peace is sustained (Novelli and Smith 2011). The UNICEF literature review and case studies in Sierra Leone, Lebanon, and Nepal (Smith, McCandless, Paulson, and Wheaton 2011; Novelli and Smith 2011) demonstrate that, among agencies and practitioners working in the education sector, the concept of peacebuilding is often unclear, its relationship to education underdeveloped, and the concept often greeted with a degree of suspicion and skepticism. In Lebanon, for example, peacebuilding was often equated with the Arab-Israeli conflict and treated with the utmost suspicion (Novelli and Smith, 2011, 24). In both Lebanon and Nepal we also encountered, on the one hand, a highly reductionist view of education s role in peacebuilding that limited it to peace education or changing minds and behavior, rather than addressing the more structural issues of governance, equal access, and quality. On the other hand we encountered the acceptance of a paradoxically broad conceptualization that essentially equates all educational activities with peacebuilding without any analytical clarity, drawing instead on a generic and well-developed rights-based discourse. For example, several informants in Sierra Leone expressed the idea that all education provision was somehow related to peacebuilding, but there was little recognition of the damage education could do by exacerbating inequality and undermining peace (see Bush and Saltarelli 2000). Interviewees across the case studies lacked a coherent vocabulary to differentiate between long-term, structural education interventions that contributed to peacebuilding (e.g., curriculum reform, reorganizing education funding to redress inequalities); short-term educational interventions that targeted particular conflict and security-related phenomena (e.g., the educational reintegration of child soldiers, refurbishing schools); and more specific thematic education interventions that supported reintegration, economic growth, social cohesion, etc., as part of broader peacebuilding interventions (e.g., technical and vocational education and training for ex-combatants). These previous studies imply that key staff working in the broad area of peacebuilding and conflict as both policymakers and practitioners rarely have sufficient knowledge of education. In contrast, education advisors and practitioners normally have a strong education background but little training in or sufficient confidence to debate the role of education in conflict and peacebuilding. There clearly is a need for greater understanding of the implications peacebuilding has for the different agencies involved in conflict-affected countries, and for a common language to discuss the components and parameters (Barnett, Kim, July

22 NOVELLI, LOPES CARDOZO, AND SMITH O Donnell, and Sitea 2007). The absence of such a language causes education and peacebuilding communities to remain in silos and results in missed opportunities for both sectors. A further tension lies in this siloed approach between the humanitarian, development, and security sectors, each of which has its own logic and agenda that intersect with education in complex ways (see Winthrop and Matsui 2013). While progress has been made in recognizing education s potential role in the humanitarian conflict and postconflict phases, it still is perceived as peripheral to the core business of shelter, food, and medical attention. This is an issue of priorities and timing, with education seen not a short-term imperative but as a long-term goal. The security sector also sees education as a marginal component that can wait until the postconflict development phase. Meanwhile, although the development sector sees education as central to its objective of helping the poor, it often is framed in terms of its economic potential (human capital) while its role in social cohesion and other broader dimensions of social justice is often underplayed. Although underpinned by somewhat different global agendas, the security and development sectors both tend to frame education s role narrowly by focusing on its market-oriented and productive outcomes, rather than on the more comprehensive sociocultural, political, and reconciliatory aspects of peacebuilding. Compounding these problems is the fact that, while it was previously thought that the humanitarian, security, and development sectors each operated in a different timeframe, they are increasingly operating simultaneously in many conflict-affected contexts. However, as the UNICEF review finds (Novelli and Smith 2011), they do not necessarily operate in a complementary manner. They also have different resources, the security sector being the most powerful, due to its links to defense and diplomacy departments. Under these conditions, better collaboration and coordination might lead one sector to dominate others, rather than the different sectoral priorities being incorporated (Novelli 2010). A disconnect between various national government departments (e.g., ministries responsible for justice, youth, gender, employment, and land rights) and between these departments and education results in an absence of cross-sector collaboration to leverage change, which under other conditions could address intersectoral issues and make education a component of a broader peacebuilding agenda. What this discussion so far highlights is the difference between the global education agenda and the distinct needs of societies that have been affected by and are emerging from conflict (Novelli, Higgins, Ugur, and Valiente 2014). The 22 Journal on Education in Emergencies

23 THE 4RS FRAMEWORK global education agendas of Education For All, Universal Primary Education, and the Millennium Development Goals, while emphasizing equity, were strongly influenced by concerns about economic productivity and efficiency. While the Education For All and Millennium Development Goals of the 1990s and 2000s underpin the relation between education and social justice as a fundamental right (Shah and Lopes Cardozo 2014, 2015; Lopes Cardozo and Shah 2016), Robertson and Dale (2013) argue that this focus on social justice has been fairly meager. They claim it has over-emphasized the distribution of access and paid too little attention to other important dimensions of social justice: Education governance frameworks therefore, both intrinsically and necessarily, have social justice implications in that they structure, and are strategically selective... of, some interests, life chances and social trajectories over others (3). However, postconflict societies may demand putting greater focus on education s potential to address inequalities and prioritize interventions that favor the promotion of social cohesion and reconciliation, along with more economic approaches. Linked to this is a disconnect between education s potential to contribute to broad societal change and narrowly defined education policies and programs. As a result of this mismatch, education policy and programs are sometimes framed within technical parameters that bypass pivotal peace-related issues in postconflict societies, including the rectification of economic, political, social, and cultural inequalities within and between groups. The conceptual language adopted recently in the Sustainable Development Goals could open up a more comprehensive and intersectoral approach, yet it remains unclear whether creating separate goals will advance or sustain segregated work within silos. 4 More comprehensive approaches require new thinking on what a sustainable peacebuilding education might look like. They require a context-sensitive approach that builds on the specific political economy and conflict dynamics of each country, and on the ways education might support broader peacebuilding goals. We recognize Bush and Saltarelli s (2000) influential report, The Two Faces of Education in Ethnic Conflict, which suggests that simply restoring the provision of education after a conflict is insufficient if the goal is to promote positive peace. While education has the potential to play such a role in postconflict societies (what they call the positive face of education), it can also do harm (the negative face; see also Smith and Vaux 2003; Smith 2010; Shah and Lopes Cardozo 2015). The 4Rs framework aims to stimulate this process of rethinking what policy and programmatic responses, and research initiatives, might look like if they move beyond a narrow technical framing of education to an approach that starts from a 4 The Sustainable Development Goals are a set of 17 global goals that have 169 targets between them. July

24 NOVELLI, LOPES CARDOZO, AND SMITH more comprehensive 4Rs-inspired conflict analysis, while simultaneously planning for future outcomes that address the interconnected dimensions of redistribution, recognition, representation, and reconciliation. We therefore now turn to the theoretical tools that have inspired our own rethinking and development of this model, acknowledging that it is not a fixed model but a process of ongoing theorybuilding that needs thematic and context-specific adaptations. A METHODOLOGICAL APPROACH INSPIRED BY SOCIAL JUSTICE AND RECONCILIATION In this section we address the idea of social justice, while also recognizing the longstanding debates in both academia and the policy field on what justice should be (Lauderdale 1998, in Zwarteveen and Boelens 2014, 146). We align ourselves with a range of scholars who aim to move beyond (1) the historical positivist (neo)liberal and utilitarian interpretations of justice (Hayek 1944; Friedman 1962; Bentham, 1981, 1988, cited in Zwarteveen and Boelens 2014, 146); and (2) the legal and uniform interpretations of justice that build on the influential work of Rawls (1971, cited in Zwarteveen and Boelens 2014). In that these theories focus mainly on abstract and universal models of redistribution to address inequality, they essentially fail to take into account the experiences and claims of marginalized groups in society (Zwarteveen and Boelens 2014, 147) and leave out a proper analysis of the social, cultural, and political conditions that underlie unequal distribution in the first place. What is needed instead is a historically informed, relational, and place-based conceptualization of justice (Zwarteveen and Boelens 2014), which in our view is what Nancy Fraser s model to a large extent offers us. Building on Fraser s (2005) work, we position the potentially transformative role education can play as inherently connected to and embedded in processes of social justice and societal transformation. Fraser, a philosopher by training who departs from but is not limited to a critical feminist perspective, asserts that a socially just society would entail parity of participation. She argues further that, to ensure participation on par with others, as full partners in social interaction (73), one should adopt the economic solution of redistributing resources and opportunities and include sociocultural remedies for better recognition and political representation. There also is a need for reconciliation processes that deal with historic and present tensions, grievances, and injustices in order to build a more sustainable peaceful society. 24 Journal on Education in Emergencies

25 THE 4RS FRAMEWORK The 4Rs model is geared toward these conflict-affected and postconflict contexts. Hence, we start with Frasers theory and adapt it to the various insights of scholars working on the relation between peacebuilding, reconciliation, and social justice (Hamber and Kelly 2004), and the relation between education and social justice (Young 2006; Connell 2012; Keddie 2012; Robertson and Dale 2013; Lopes Cardozo and Shah 2016). Although focused primarily on Western education systems in less conflict-prone environments, Keddie (2012) attempts to apply Fraser s three-dimensional model to educational injustice. While acknowledging some of the critiques and debates around Fraser s work, Keddie convincingly claims that Fraser s model should not be offered as an ideal of justice that is static and uncomplicated but rather as a productive lens for thinking about and addressing some of the key ways in which different dimensions of injustice are currently hindering the schooling participation, engagement and outcomes of marginalised students (15). Tikly and Barrett (2011) argue further that, in developing contexts, a social justice approach that draws on the work of Fraser and Sen can provide a fuller rationale for a policy focus on education quality than that provided by a human capital approach with its emphasis on economic growth or by the existing human rights approach with its emphasis on the role of the state in guaranteeing basic rights (3-4). For the purpose of our analysis, we apply these insights to studying injustice in and through education in conflictaffected regions, where sociocultural, political, and economic inequalities are often the root causes of tension and violence. It is important to note that, in keeping with Fraser s line of thought, while the dimensions of the 4Rs are separated for analytical purposes, they actually are closely interlinked. We also need to acknowledge how internal relations between these Rs can be reinforcing or conflictive. For example, recognizing formerly excluded ethnic languages in education and redistributing resources to train teachers and develop material to enhance this process could lead to greater representation of ethnic minority graduates in decision-making positions at the school governance level or later in political positions. However, opening up to diverse languages also might hinder the reconciliation process, as some minority languages might be included as a language of instruction while others are not, thus creating resentment among various groups of students. What does this theoretical inspiration about social justice, inspired primarily by Fraser, imply for our methodological choices? For one thing, we feel that an interdisciplinary and intersectoral approach to the research is necessary precisely because education and peacebuilding are affected from both within and outside the sector. Thus we need to move beyond educationism and the idea that we July

26 NOVELLI, LOPES CARDOZO, AND SMITH can understand education from within itself and recognize that education policy, systems, programs, and practices are embedded in complex local, national, and global political economies that both shape and are shaped by this relationship (Dale 2005; Dale and Robertson 2009; Robertson and Dale 2014). We also need a methodological approach that neither reifies nor privileges local, national, or global geographic scales and instead seeks to develop a framework for understanding the complex relationships between scales and interrogates multiscalar relationships. In other words, we want to avoid drifting into either modernization theories blindness toward exogenous factors or dependency theories often equally myopic avoidance of endogenous factors. This requires tracing policies, practices, and power across local, national, and global actors and factors to understand education and peacebuilding activities more fully in short, we need a multiscalar approach (Novelli, Higgins, Ugur, and Valiente 2014). However, this approach should not be applied only to research initiatives, as recurrent messages in the literature point to the failure of state-centric approaches by international actors to connect to the agency of local actors within civil society and in sub-national contexts. This failure limits or undermines the ability to capitalize on the knowledge and peacebuilding practices of local actors, and to respond to their educational needs and aspirations. It also creates a disjuncture between a rigidly supplied education and flexible and varied community demands for educational provision (Novelli, Higgins, Ugur, and Valiente 2014). Also needed is a highly sensitive methodological approach that is attentive to the particular contexts in which the research is taking place. This requires that the research be located within a well-developed cultural, political, economic, and conflict analysis of the particular places and spaces being examined, and a recognition of culture as centrally embedded in these analyses. We refer here to issues related to ethnicity, gender, cultural and religious heritage, and civilizational issues (see Robertson and Dale 2013). This leads us to adopt a critical, cultural, political economy approach (Jessop 2005; Sum and Jessop 2013; Robertson and Dale 2014; Lopes Cardozo and Shah 2016), which seeks to bridge materialist and poststructuralist approaches to understanding the political economy. It recognizes the complex interplay between language/culture and the interconnected materialities of economics and politics within wider social formations. We believe that such a critical, cultural, political economy analysis of education (Robertson and Dale 2014) can provide a comprehensive framework to help understand first, how the relationship between education and peacebuilding is articulated discursively and materially through social relations, experiences, and practices (the cultural); second, the ways education and peacebuilding fit into relations of 26 Journal on Education in Emergencies

27 THE 4RS FRAMEWORK production, distribution, and exchange in society (the economic); and third, the way an agenda promoting education s links to peacebuilding has been determined and subsequently governed (the political). Finally, we want to recognize that, in their application in policy, programming, or research, the concepts of the 4Rs model must be translated in particular conflict contexts into local understanding of redistribution, recognition, representation, and reconciliation. Boaventura de Sousa Santos (2007) has noted that there can be no global social justice without global cognitive justice, which requires a process of decolonizing the social sciences by opening up alternative knowledges, approaches, and paradigms that emerge from the global south. He calls for a process of translation that brings alternative understandings of social issues and problems into dialogue through a process of translation. Santos poses the problem of how different groups with different histories, objectives, and trajectories can unite around certain common issues. He talks of the possibility of drawing together at the global level the concept of human rights and the Hindu and Islamic concepts of human dignity; Western strategies of development and Ghandi s swadeshi (Santos 1998); Western philosophy and African oral sagesse; modern democracy and traditional authorities; the indigenous movement and the ecological movement, etc. The task of a politics of translation, then, is to facilitate communication between different subjects. This entails recognizing the other as a producer of knowledge while bearing in mind a sociology of absences that is, an understanding of the hierarchy of the available hegemonic and sometimes silenced counter-hegemonic discourses; a move from decontextualized absolute knowledge to forms of contextualized knowledge; and a focus on the duality between conformist action and rebellious action, particularly the attempt to reconstruct the idea and practice of emancipatory social transformation (Santos 1998, 133). While knowledge-as-regulation has been (and often still is) the dominant form, Santos encourages us to reinvent knowledge-as-emancipation, and the need for alternative thinking of alternatives (129). In practical terms, the process of translation and alternative thinking of alternatives that the 4Rs model calls for requires researchers, policy-makers, and practitioners to engage with existing scholarly work in each context we are working in, and with local academics, social activists, and practitioners, to get a sense of how our conceptions of social justice and reconciliation align with and support other conceptions of human dignity and do not and the forms this takes. While interdisciplinary and international collaborations between researchers, policy-makers, and practitioners logically follow from this line of thinking, we recognize the remaining challenges in working in intercultural and July

28 NOVELLI, LOPES CARDOZO, AND SMITH multilingual teams, and in contexts with unequal access to resources, including online working and communication facilities. Our hope is that working with the 4Rs model will stimulate serious engagement and a process of (self-)reflection, and that it will promote constant collaborative decision-making aimed toward socially just studies, policies, and programs. EDUCATION FOR SUSTAINABLE PEACEBUILDING: THE 4RS FRAMEWORK Our analytical framework contends (1) that sustainable peacebuilding is dependent on societal transformation; (2) that social sectors (including education) play a crucial role in such transformation; and (3) that the transformative processes involve not just the three Rs as suggested by Fraser, but a combination of the four Rs discussed above and shown in Figure 1. In this visualization of the 4Rs model, we emphasize the porous boundaries and interconnections between the Rs, as depicted by the arrows that connect the 4Rs segments. The outer white arrow signals how the 3Rs Fraser (1995, 2005) developed require a thorough analysis of the various drivers of injustice and, in contrast, how the positive face of education could be supported by addressing such concerns. The gray arrow, visualized in relation to the reconciliation segment, requires us to engage with an analysis of the various legacies of violent conflict, and at the same time to envision ways that education governance, content, and pedagogy can support education s potential contributions to transitional justice, healing, and trust-building. Although we find it useful for analytical purposes to separate out these four dimensions, we want to recognize from the outset that they are parts of a complex whole that needs to be understood relationally. As Figure 1 illustrates, the first R of redistribution provides a range of remedies for social injustice caused by unequal distribution of resources, exclusive participation in economic structures, and a lack of equal social (educational, health, employment, etc.) opportunities. The second R of recognition entails possible solutions to injustice that have to do with status inequalities, which prevent some people from having equal or full interaction in institutionalized cultural hierarchies. This is often related to there being little acceptance or space for cultural, ethnic, linguistic, racial, gender, age, or other types of diversity. The third R, for representation, leads us to analyze the (absence of) transformative politics on multiple scales global, national, local which lead to citizens unequal participation in decision-making or claim-making processes (Fraser 2005). The fourth R is for reconciliation, and it takes us beyond Fraser s work into a process 28 Journal on Education in Emergencies

29 THE 4RS FRAMEWORK that is crucial for postconflict societies in order to prevent a relapse into conflict. It incorporates education s role in dealing with the past and with historic memory, truth and reparations, transitional justice processes, issues related to bringing communities together, processes of forgiving and healing, and the broader process of social and psychosocial healing (see Hamber 2007, 2009). Figure 1: Sustainable Peacebuilding in Education: The 4Rs Analytical Framework Equitable access to education Equitable distribution of resources Outcomes (qualifications, employment opportunities) Analysis of education reforms/ policies to see if they are redistributive Language of instruction polices Recognition of cultural diversity through curriculum Place of religious and cultural identity in the education system Citizenship and civic education as a means of state-building Addressing historic and contemporary economic, political, and cultural injustice Analysis of how education strengthens/weakens social cohesion Teaching about the past and its relevance to the present and future Levels of trust vertical (in schools and the education system) and horizontal (between groups) Extent to which education policy/ reforms involve stakeholders participation (local, national, global) Analysis of political control/representation through administration of education School governance, involvement in decision-making (teachers, parents, students) Extent to which the education system supports fundamental freedoms In simple terms, Fraser s 3Rs help us analyze and understand the different dimensions of the drivers of conflict in various contexts and in relation to education, while the fourth R of reconciliation helps us explore the legacies of conflict in relation to education. Addressing both the drivers of conflict and the legacies of conflict is a complex process, but one that is crucial for the promotion of sustainable peace. In our effort to develop an analytical model of peace with social justice that is relevant for the analysis of peacebuilding and education in conflict-affected contexts, we find that a relational dimension (Hamber and Kelly 2004) of reconciliation is indispensable. We argue that, as we add reconciliation to an July

30 NOVELLI, LOPES CARDOZO, AND SMITH existing framework, we need to elaborate a bit more on how we can understand and include reconciliation in a sustainable peacebuilding model and how it connects to the other three Rs. We draw here on a useful and dynamic definition developed by Hamber and Kelly (2004), who see reconciliation in postconflict environments as a process of addressing conflictual and fractured relationships through voluntary acts that cannot be imposed (3-4). Reconciliation also should be considered a paradoxical process, as it promotes an encounter between the open expression of the painful past on the one hand while it seeks a long-term, interdependent future on the other (Lederach 1997, cited in Hamber and Kelly 2004). Hamber and Kelly (2004) further define five interconnected strands of reconciliation: The development of a shared vision of an interdependent and fair society Acknowledging and dealing with the past The need to build positive relationships that address issues of trust, prejudice and intolerance The need for significant cultural and attitudinal change The need for substantial social, economic, and political change Education s potential role lies in providing or supporting what these authors call mechanisms for justice, healing, restitution or reparation, and restoration. According to Hamber and Kelly (2004), individuals and institutions can acknowledge their role in historic conflicts, and by doing so learn to avoid a relapse into conflict. A concrete example of education s role in reconciliation processes and in dealing with a conflictive past is the teaching of history. We also recognize Hamber and Kelly s (2004) warning that the concept of reconciliation is always influenced by people s underlying assumptions or ideologies religious, political, economic, or other. Hence, in our own understanding of reconciliation as part of our analytical framework, we recognize the need to develop contextualized, locally defined, and historically informed understandings of what reconciliation could/should mean in the very different contexts under study. Furthermore, while it is important to bring people from different and even opposing social groups together, either through 30 Journal on Education in Emergencies

31 THE 4RS FRAMEWORK formal integrated schooling or non-formal programs, it is important to move beyond a narrow interpretation of Allport s Contact Hypothesis and allow for meaningful long-term encounters and reflection. 5 Simply getting together to shake hands and share food is to suggest that conflict is driven (only) by interpersonal animosities rather than (also) by structural grievances and inequalities. Finally, we recognize that there is significant tension between Nancy Fraser s 3Rs of social justice and the fourth R of reconciliation. While the former seeks to identify and reduce the drivers of conflict, reconciliation is much more concerned with dealing with the aftermath, or legacies of conflict, and bringing people and communities together. The balance between policies that promote social justice (and therefore address the drivers of conflict) and those that promote reconciliation (and address the legacies of conflict) is therefore a political decision that needs to be worked out on the ground by the key stakeholders in each particular context. APPLYING THE 4RS TO ANALYZE THE RELATION BETWEEN EDUCATION AND PEACEBUILDING So what does this analytical framework mean in terms of examining the relationships between education and peacebuilding processes, whether in research projects or when designing or reviewing policy-related or programmatic work? Sustainable peacebuilding should not be conceptualized just as a means to education (access) but also in and through education. It should consider how teaching and learning processes and outcomes reproduce certain (socioeconomic, cultural, and political) inequalities (Keddie 2012) and thus can stand in the way, or reinforce, processes of reconciliation and foster education s negative, or positive, face. Hence, we now explore how we can apply the 4Rs analytical model to look at specific and contextualized educational problems, along with possible responses in conflict-affected situations. Our aim is to demonstrate the usefulness of the model for (academic) analytical relevance and its more practical utility in the planning and evaluation of concrete initiatives. We also see the 4Rs model as a possible approach to design and structure (research, programmatic) projects, whereby starting from a comprehensive 4Rs-inspired context-and-conflict analysis informs the choices made. The 4Rs framework also has been applied to analyze and examine the way specific interventions positively or negatively impact sustainable peace outcomes on various fronts. To do justice 5 The contact hypothesis has been described as one of the best ways to improve relations among groups that are experiencing conflict (Allport 1954). July

32 NOVELLI, LOPES CARDOZO, AND SMITH to education s full potential, the model aims to move away from narrow technical approaches to understanding, designing, and implementing education in conflictaffected regions, and toward a model that allows for the exploration of and positive engagement with a wider range of conflict drivers and legacies. Table 1: Applying the Analytical Framework to Understand Education s Role in Peacebuilding To what extent is education contributing to sustainable peacebuilding (4Rs)? Potential indicators for a mixed-methods approach Redistribution (addressing inequalities) Quantitative analysis of existing data to examine vertical and horizontal inequalities relevant to education inputs, resources, and outcomes Analysis of macro education reforms or policies to see if they are redistributive; for example, the impact of decentralization, privatization, and how they impact different groups and affect conflict dynamics Recognition (respecting difference) Language of instruction polices Recognition of cultural diversity through curriculum Place of religious identity and religious diversity in teaching practices (Re)production of gendered relations and norms in the education system Citizenship, civic, sexuality, and history education in relation to statebuilding Representation (ensuring participation) The extent to which education policy and reforms are produced through participation (local, national, global) Analysis of political control and representation through the administration of education School governance, school-based management, involvement in decision-making (teachers, parents, students, civil society) The extent to which education system supports fundamental freedoms, including equal gender representation 32 Journal on Education in Emergencies

33 THE 4RS FRAMEWORK (Table 1 cont.) To what extent is education contributing to sustainable peacebuilding (4Rs)? Potential indicators for a mixed-methods approach Reconciliation (dealing with the legacies of the conflict) The extent to which the historical and contemporary economic, political, and sociocultural injustices that underpin conflict are redressed in/through education (e.g., quota systems, school relocation, textbooks, teacher allocation) Analysis of how education contributes to integration and segregation (social cohesion, shared or separate institutions) Teaching about the past and its relevance to the present and future (Dis)connection of educational activities to the work of truth and reconciliation committees, when available Levels of trust vertical (trust in schools and the education system) and horizontal (trust between different identity-based groups) A number of important aspects emerge when exploring the four interrelated Rs. An important aspect of redistribution (not limited to this dimension) is all students having equal access to a safe journey to and through their learning environment. Within education, the inclusion of all students regardless of age, gender, sexuality, religion, ethnicity, race, language, class, etc. means including formerly marginalized or disadvantaged groups. This aspect is also connected to reconciliation. The affirmation and recognition of learners diversity and everyone s learning needs in educational processes, structures, and content can be defined as curricular justice (Connell 2012). This aspect of recognition is strongly related to the redistributive aspect of equal opportunities and outcomes for children and youth of different groups in society. The structure and content that feed into pedagogical processes are again connected to both reconciliation (e.g., if/how history is taught or if attitudinal change is part of an educational initiative) and representation (e.g., whether learners are made aware of their various rights and responsibilities as citizens, and if/how/why [certain] political and conflict-related issues are discussed/negated). Issues around representation extend further into the actual equitable participation of various stakeholders, including teachers, students, youth, parents, and community members of all genders at the grassroots level. The actual decision-making power is often related to the allocation, use, and (re)distribution of human and material resources (Young 2006; Robertson and Dale 2013). Our research in Myanmar (Higgins, Maber, Lopes Cardozo, and Shah 2016) provides an example of how the 4Rs framework opened up new insights into the ways education-sector reform, educational governance mechanisms, and pedagogical practices interrelate in complex and often troublesome ways with July

34 NOVELLI, LOPES CARDOZO, AND SMITH ongoing inequalities and tensions in Burmese society. A central issue in the current landscape of Myanmar is the ongoing peace negotiations between the government and multiple armed ethnic groups, which are as yet unresolved after six decades of fighting. Education is not an explicit component of the Nationwide Ceasefire Agreements, but it is an important aspect of the peace dialogue and is perceived to be a key grievance of many of the armed ethnic groups, civil society organizations, and minority groups. While current education reform is deemed vital to securing peace dividends through improved service delivery and a renewed focus on inclusion and equal provision, peacebuilding seems to be only an implicit part of broader discourses of social inclusion, equity, and improved access for those traditionally neglected by the state; it is not explicitly mentioned in education reform discourse. Our analysis of drafts of the Comprehensive Education Sector Reform, National Education Strategy Plan, and Education Law suggests that peacebuilding is everywhere and nowhere: everywhere in the sense that there seems to be a recognition of the need to place education reform in the context of real inequalities and frustrations, but nowhere as peacebuilding logic or language. In that sense, while reconciliation is to some extent part of the latest government discourse, many education stakeholders including teachers, students, and civil society actors feel that reconciliation efforts are hardly taking place, especially considering ongoing tensions and struggles. Teaching history was brought forward by students and teachers alike as an area that could foster reconciliation and social cohesion, but teachers felt too constrained by existing curricula to allow this potential positive face of education to be presented. In Myanmar, then, the reform and policy direction of the 4Rs might do more harm than good by failing to address the root causes of the conflict in the first place: a lack of fair (re)distribution of resources and opportunities; recognition of citizens various linguistic and cultural needs; sincere representation and a participatory process that not only informs but engages with oppositional and minority perspectives to enable the first steps forward in addressing the grievances expressed through and inflicted by education through reconciliation. Genderbased forms of inequity are either absent from the reform process or tend to focus only on quantitative parity in school enrollments and completion. In fact, little attention has been given to the gendered forms of bias, discrimination, and (structural and indirect) violence that the education system and structures have imposed on learners and communities for decades. They also are not appropriately considered and addressed in current reform efforts. 34 Journal on Education in Emergencies

35 THE 4RS FRAMEWORK Our findings further highlight another key tension within this reform package, this one between the aim to deliver quick, visible peace dividends and the desire to take a systematic, evidence-based approach to the education sector as a whole. We observed a possible trade-off between the Rs, in the sense that a focus on redistributing educational resources and reducing access-related barriers to schooling may work against the goals of recognizing a plurality of viewpoints and actors (e.g., regarding language of instruction). Despite efforts to ensure inclusivity, consultation has been fairly unrepresentative and many key national stakeholders feel sidelined. There is an uneasy tension between a state that has expressed its intention to address issues of redistribution, representation, and recognition by decentralizing a strongly centralized system, and the connected process of convergence that raises questions about whether decentralization would limit opportunities for citizens, particularly ethnic minority groups, to vocalize and represent their education interests, at least in the short term. Finally, recent efforts to revise the formal school curriculum in Myanmar again do not explicitly refer to peacebuilding processes not as a literal translation but as the broad conceptualization we employ in this framework. However, specific components of the envisaged content, including fewer references to the military in social studies, indicate a commitment to delegitimize violence. Nonetheless, textbooks written primarily in the majority Burmese language and uncertainty about how the curriculum will include 20 percent local content point out the curriculum s limitations in addressing pressing issues of social justice, which is linked to a lack of recognition of the diverse cultures and the representation of minority groups. Hence, coming back to the five interconnected strands of reconciliation (Hamber and Kelly 2004), our analysis of Myanmar s ongoing reform efforts, and even more so the educational realities, show little development of a shared vision for an interdependent and fair society, and a minimal attempt to deal with the past in and through education. At the same time, building positive relationships and trust, and significant cultural and attitudinal change, were only observed in non-formal forms of education led mostly by civil society. We argued in our research on education s role in peacebuilding in Myanmar, in line with Hamber and Kelly s fifth strand of reconciliation, that education alone is not in a position to create peace; it must be integrated into a process of substantial social, economic, and political change. At the same time, a narrow approach to the technical and economic functions of education from a human capital perspective severely limits its potential positive contribution to peacebuilding. In contrast, a more holistic approach to education governance, content, and pedagogy that addresses all 4Rs July

36 NOVELLI, LOPES CARDOZO, AND SMITH would allow younger generations to support their sociocultural, political, and economic agency for peace (Lopes Cardozo, Higgins, and Le Mat 2016). Overall, applying the 4Rs framework to our analysis of the data and findings in the case of Myanmar (and a range of other cases referred to below) illustrates the closely interrelated connections, and often the contested nature, between the four dimensions of redistribution, recognition, representation, and reconciliation. While diagnosis is no guarantee for a cure, the analysis and recommendations that emerged from the research help to challenge education reforms currently taking place in Myanmar. These reforms are supported by both national and international actors that bypassed the nuanced and complex issues raised and instead reproduced a generic education menu that appears ill-suited to the contexts and scale of the conflicts and education challenges in these countries. CONCLUSION: THEORY-BUILDING IN PROGRESS In this article we have shared the 4Rs analytical framework, which calls for a normative peace with social justice and reconciliation approach to education systems affected by violent conflict. While aspects of the model are potentially relevant across different contexts, it must be tailored to the specific needs of each area of research or intervention, as we have concluded in the recent application of the model in Kenya (Smith, Marks, Novelli, Valiente, and Scandurra 2016); Myanmar, Pakistan, South Sudan, South Africa, and Uganda (Datzberger, McCully, and Smith 2015; Smith, Datzberger, and McCully 2016a; Novelli, Daoust, Selby, Valiente, Scandurra, Deng Kuol, and Salter 2016; Higgins, Maber, Lopes Cardozo, and Shah 2016); and Sri Lanka (Duncan and Lopes Cardozo 2017). This will allow researchers and practitioners alike to produce high-quality, relevant understanding of the challenges, roles, and possibilities of education s contribution to sustainable peacebuilding. In that sense, the 4Rs approach is a heuristic device that can spark a dialogue among key stakeholders, a framework that will enable us to ask the right questions and reflect on the dilemmas and contradictions inherent in working to support education s positive role in peacebuilding. In so doing, we hope to refine, develop, sharpen, and transform the framework so it can more accurately reflect the combined knowledge that emerges from the ongoing research process. In that sense, we approach theory-making as a nonstatic process that is informed and reshaped through empirical fieldwork and findings hence this framework as theory-building in progress. We welcome feedback and suggestions from those interested in exploring the usefulness of this 36 Journal on Education in Emergencies

37 THE 4RS FRAMEWORK proposed model (in programming, policy development, and academic studies) to further improve our collective understanding of the complex relationship between education and the processes of sustainable peacebuilding. ACKNOWLEDGMENTS We are grateful to the comments and valuable input provided on the first drafts of this text by the three respective teams of the Research Consortium on Education and Peacebuilding situated at the University of Amsterdam, the University of Sussex, and Ulster University. We also thank UNICEF for the collaboration and the opportunity to share and discuss this framework in numerous meetings at headquarters and in country offices. We are also grateful for the feedback of the anonymous reviewers and the editorial team of the Journal on Education in Emergencies, which has contributed positively to the end result published here. REFERENCES Allport, Gordon The Nature of Prejudice. Reading, MA: Addison-Wesley. Barnett, Michael, Hunjoon Kim, Madalene O Donnell, and Laura Sitea Peacebuilding: What Is in a Name? Global Governance: A Review of Multilateralism and International Organizations 13 (1): Bush, Kenneth D., and Diana Saltarelli, eds The Two Faces of Education in Ethnic Conflict: Towards a Peacebuilding Education for Children. Florence, Italy: UNICEF Innocenti Research Centre. Castañeda, Carla How Liberal Peacebuilding May be Failing Sierra Leone. Review of African Political Economy 36 (120): Connell, Raewyn Just Education. Journal of Education Policy 27 (5): Dale, Roger Globalisation, Knowledge Economy and Comparative Education. Comparative Education 41 (2): July

38 NOVELLI, LOPES CARDOZO, AND SMITH Dale, Roger, and Susan L. Robertson Beyond Methodological Isms in Comparative Education in an Era of Globalisation. In Handbook on Comparative Education, edited by Andreas M. Kazamias and Robert Cowen. Dordrecht: Springer, Datzberger, Simone, Alan McCully, and Alan Smith Education and Peacebuilding in Uganda. New York: UNICEF. Denney, Lisa Reducing Poverty with Teargas and Batons: The Security- Development Nexus in Sierra Leone. African Affairs 110 (439): Duffield, Mark Post Modern Conflict: Warlords, Post Adjustment States and Private Protection. Civil Wars 1 (1): Duncan, Ross, and Mieke T. A. Lopes Cardozo Reclaiming Reconciliation through Community Education for the Muslims and Tamils of Post-War Jaffna, Sri Lanka. Research in Comparative and International Education 12(1): Fraser, Nancy From Redistribution to Recognition? Dilemmas of Justice in a Post-Socialist Age. New Left Review 1 (212): Fraser, Nancy Reframing Justice in a Globalized World. New Left Review 36: Galtung, Johan Three Realistic Approaches to Peace: Peacekeeping, Peacemaking and Peacebuilding. Impact of Science on Society 26 (1-2): Galtung, Johan Cultural Violence. Journal of Peace Research 27 (3): Hamber, Brandon Forgiveness and Reconciliation: Paradise Lost or Pragmatism? Peace and Conflict 13 (1): Hamber, Brandon Transforming Societies after Political Violence: Truth, Reconciliation, and Mental Health. New York: Springer. Hamber, Brandon, and Gráinne Kelly A Working Definition of Reconciliation. com/pubs_papers.htm. 38 Journal on Education in Emergencies

39 THE 4RS FRAMEWORK Heathershaw, John Unpacking the Liberal Peace: The Dividing and Merging of Peacebuilding Discourses. Millennium: Journal of International Studies 3: Higgins, Sean, Elizabeth Maber, Mieke T. A. Lopes Cardozo, and Ritesh Shah Education and Peacebuilding in Myanmar. Research Consortium Education and Peacebuilding, University of Amsterdam; summary of report is available online, outputs-research-consortium/. Jessop, Bob Critical Realism and the Strategic-Relational Approach. New Formations 56: Keddie, Amanda Schooling and Social Justice through the Lenses of Nancy Fraser. Critical Studies in Education 53 (3): Lederach, John Paul Preparing for Peace: Conflict Transformation across Cultures. Syracuse, NY: Syracuse University Press. Lederach, John Paul Building Peace: Sustainable Reconciliation in Divided Societies. Washington, DC: U.S. Institute of Peace Press. Lopes Cardozo, Mieke T. A., Sean Higgins, and Marielle L. J. Le Mat Youth Agency and Peacebuilding: An Analysis of the Role of Formal and Non-Formal Education Synthesis Report on Findings from Myanmar, Pakistan, South Africa and Uganda. Research Consortium Education and Peacebuilding, University of Amsterdam. files.wordpress.com/2016/06/youth-agency-synthesis-report-final16.pdf. Lopes Cardozo, Mieke T. A., and Celine M. Q. Hoeks Losing Ground: A Critical Analysis of Teachers Agency for Peacebuilding Education in Sri Lanka. Journal of Peace Education 12 (1): Lopes Cardozo, Mieke T. A., and Ritesh Shah A Conceptual Framework to Analyse the Multiscalar Politics of Education for Sustainable Peacebuilding. Comparative Education 52 (4): McCandless, Erin Peace Dividends and Beyond: Contributions of Administrative and Social Services to Peacebuilding. New York: United Nations Peacebuilding Support Office. July

40 NOVELLI, LOPES CARDOZO, AND SMITH Novelli, Mario The New Geopolitics of Educational Aid: from Cold Wars to Holy Wars? International Journal of Educational Development 30: Novelli, Mario. 2011a. The Role of Education in Peacebuilding: Methodological Framework for Three Country Case Studies. New York: UNICEF. Novelli, Mario. 2011b. The Role of Education in Peacebuilding: Sierra Leone Case Study. New York: UNICEF. Novelli, Mario The Merging of Security and Development in the Education Sector: Discourses, Programmes and Effects (Discussion paper for the Open Society Foundation, Education Support Programme). London: OSI. Novelli, Mario, Gabrielle Daoust, Jan Selby, Oscar Valiente, Rosario Scandura, Luka Biong Deng Kuol, and Emma Salter Education Sector Governance, Inequality, Conflict and Peacebuilding in South Sudan, (Research Report UNICEF Eastern and Southern Regional Office). New York: UNICEF. Novelli, Mario, Sean Higgins, Mehmet Ugur, and Oscar Valiente Rigorous Literature Review of the Political Economy of Education Systems in Conflict Affected States. London: DFID. Novelli, Mario, and Mieke T. A. Lopes Cardozo Conflict, Education and the Global South: New Critical Directions. International Journal of Educational Development 28 (4): Novelli, Mario, and Mieke T. A. Lopes Cardozo Globalizing Educational Interventions in Zones of Conflict: The Role of Dutch Aid to Education and Conflict. In Global Education Policy and International Development: New Agendas, Issues and Policies, edited by Antoni Verger, Mario Novelli, and H. Kosar Altinyelken. London: Continuum, Novelli, Mario, and Alan Smith The Role of Education in Peacebuilding: A Synthesis Report of Findings from Lebanon, Nepal and Sierra Leone. UNICEF: New York. Paris, Roland At War s End: Building Peace after Civil Conflict. Cambridge, UK: Cambridge University Press. 40 Journal on Education in Emergencies

41 THE 4RS FRAMEWORK Paris, Roland Saving Liberal Peacebuilding. Review of International Studies 36 (2): Pugh, Michael, Neil Cooper, and Mandy Turner Whose Peace? Critical Perspectives on the Political Economy of Peacebuilding. London: Palgrave Macmillan. Research Consortium on Education and Peacebuilding Introducing the Research: Research Briefing 1. com/rp/research-consortium-education-and-peacebuilding/. Richmond, Oliver, Annika Björkdahl, and Stefanie Kappler The Emerging EU Peacebuilding Framework: Confirming or Transcending Liberal Peacebuilding? Cambridge Review of International Affairs 24 (3): Richmond, Oliver, and Annika Mitchell Hybrid Forms of Peace: From Everyday Agency to Post-Liberalism. London: Palgrave Macmillan. Robertson, Susan L., and Roger Dale The Social Justice Implications of Privatisation in Education Governance Frameworks: A Relational Account. Oxford Review of Education 39 (4): Robertson, Susan L., and Roger Dale Toward a Critical Cultural Political Economy Account of the Globalising of Education. Paper presented at the 55th Comparative and International Education Society Conference, Toronto, March Santos, Boaventura de Sousa Oppositional Postmodernism and Globalizations. Law and Social Inquiry 23 (1): Santos, Boaventura de Sousa Cognitive Justice in a Global World: Prudent Knowledges for a Decent Life. Lanham, MD: Rowman and Littlefield, Lexington Books. Selby, Jan The Myth of Liberal Peace-Building. Conflict, Security & Development 13 (1): July

42 NOVELLI, LOPES CARDOZO, AND SMITH Shah, Ritesh, and Mieke T. A. Lopes Cardozo The Politics of Education in Emergencies and Conflict. In Education and International Development: An Introduction, edited by T. McCowan and E. Unterhalter. London: Bloomsbury Academic, Shah, Ritesh, and Mieke T. A. Lopes Cardozo Building Peace and Social Justice through Education in Aceh, Indonesia? International Journal for Educational Development 38: Smith, Alan Education in the 21st Century: Conflict, Reconstruction and Reconciliation. Compare, Journal of the British Association for International and Comparative Education 36 (4): Smith, Alan The Influence of Education on Conflict and Peace Building. (Paper prepared for the Education for All Global Monitoring Report 2011, The Hidden Crisis: Armed Conflict and Education ). Paris: UNESCO. Smith, Alan Contemporary Challenges for Education in Conflict Affected Countries. Journal of International and Comparative Education (3) 1: Smith, Alan, Simone Datzberger, and Alan McCully The Integration of Education and Peacebuilding: Synthesis Report on Myanmar, Pakistan, South Africa and India. New York: UNICEF. Smith, Alan, Caroline Marks, Mario Novelli, Oscar Valiente, and Rosario Scandurra The Links between Equity, Governance, Education and Peacebuilding in Kenya. Nairobi: UNICEF. Smith, Alan, Erin McCandless, Julia Paulson, and Wendy Wheaton Education and Peacebuilding in Post-Conflict Contexts: A Literature Review. New York: UNICEF. Smith, Alan, and Tony Vaux Education, Conflict and International Development. London: DfID. Sum, Ngai-Ling, and Bob Jessop Towards a Cultural Political Economy: Putting Culture in Its Place in Political Economy. Cheltenham, UK: Edward Elgar. 42 Journal on Education in Emergencies

43 THE 4RS FRAMEWORK Tikly, Leon, and Angeline M. Barrett Social Justice, Capabilities and the Quality of Education in Low Income Countries. International Journal of Educational Development 31 (1): Vaux, Tony UNICEF Research Project on the Role of Education in Peacebuilding in Crisis and Post-Conflict Contexts: Nepal Case Study. New York: UNICEF. Winthrop, Rebecca, and Elena Matsui A New Agenda for Education in Fragile States (Working paper No. 10). Washington, DC: Center for Universal Education at the Brookings Institution. Young, Iris Marion Responsibility and Global Justice: A Social Connection Model. Social Philosophy & Policy Foundation 23 (1): Zakharia, Zeena UNICEF Research Project on the Role of Education in Peacebuilding in Crisis and Post-Conflict Contexts: Lebanon Case Study. New York: United Nations Children s Fund. Zwarteveen, Margreet Z., and Rutgerd Boelens Defining, Researching and Struggling for Water Justice: Some Conceptual Building Blocks for Research and Action. Water International 39 (2): July

44 CAN TEACHER-TRAINING PROGRAMS INFLUENCE GENDER NORMS? MIXED- METHODS EXPERIMENTAL EVIDENCE FROM NORTHERN UGANDA Marjorie Chinen, Andrea Coombes, Thomas De Hoop, Rosa Castro-Zarzur, and Mohammed Elmeski ABSTRACT This mixed-methods cluster-randomized controlled trial examines the impact of a teacher-training program that aimed to promote positive gender socialization in the conflict-affected region of Karamoja, Uganda. The theory of change suggests that the education system and teachers can play critical roles in promoting positive gender roles and gender equality, which has important implications for peacebuilding. Our study found evidence that the program positively influenced teachers knowledge about the difference between gender and sex, and their attitudes toward gender roles and gender identity. We found no quantitative evidence for any short-term change in teachers practices as a result of the program, nor did we find quantitative evidence of effects from a complementary, randomly assigned text-message intervention meant to reinforce the information delivered during the training. Qualitative research suggested that, while teachers adopted basic practices taught in the training, they were unready or unable to adopt more complex practices. The main implication is that training can influence teachers knowledge and attitudes on gender equality, but traditional gender norms can be a barrier to changing behavior in the short term. A further implication is the importance of involving the community to create enabling environments in which new ideas about gender equality can be accepted and translated into practice. Received August 30, 2016; revised February 27, 2017, and May 26, 2017; accepted June 7, 2017; electronically published July Journal on Education in Emergencies, Vol. 3, No. 1 Copyright 2017 by the Inter-Agency Network for Education in Emergencies (INEE). All rights reserved. ISSN Journal on Education in Emergencies

45 CAN TEACHER-TRAINING PROGRAMS INFLUENCE GENDER NORMS? INTRODUCTION Gender equality is a fundamental human right. When women and men do not have equal access to resources or equal opportunities, there are direct economic and social costs. These costs largely affect women, and also have consequences for their children, communities, and countries. For instance, disadvantages in education translate into a lack of skills and limited access to opportunities in the labor market, which in turn affect social progress (Sustainable Development Goals 2015). The education system not only has the potential to build children s capacities, it also can play a vital role in shaping children s understanding of gender roles and stereotypes and in internalizing positive gender norms during childhood and into adolescence. Conversely, education that legitimizes harmful gender stereotypes and provides inequitable services, biased textbooks, and biased teaching methods can reinforce exclusion and stereotypes and threaten access to education and education quality, thereby undermining children s ability to contribute to peacebuilding. Although research suggests that education can contribute to gender equality in conflict-affected environments and fragile states (Baranyi and Powell 2005, 2; Winthrop and Kirk 2008, 647), there is only limited evidence of what works to promote gender equality in education in conflict-affected settings. To address this evidence gap, more rigorous experimental and quasi-experimental evidence is needed on the topic. Furthermore, programs that aim to improve gender equality often focus on results that are hard to quantify, such as gender norms and women s empowerment (Burde, Guven, Kelcey, Lahmann, and Al-Abbadi 2015, 3). Therefore, it is crucial to supplement quantitative research with in-depth qualitative research in order to gain a better understanding of how programs work and affect these outcomes. This paper helps to reduce this knowledge gap by focusing on the impact of an eight-month gender-socialization training program for teachers in Karamoja, Uganda. We conducted a cluster-randomized controlled trial (RCT) that randomly assigned 35 schools to a control group, 35 schools to the UNICEF-supported teacher-training program (Treatment 1), and 35 schools to the same program but with the addition of complementary text messages that reinforced training information (Treatment 2). In line with our theory of change, we estimated the impact the program would have on teachers knowledge, attitudes, and selfreported practices around gender equality. We administered structured teacher surveys during baseline (March 2015) and endline (November 2015) data July

46 CHINEN, COOMBES, DE HOOP, CASTRO-ZARZUR, AND ELMESKI collection. To increase our understanding of the program s intangible results, we supplemented the quantitative analysis with the collection and analysis of qualitative data. We found evidence that the program had positive effects on teachers knowledge about the difference between gender and sex, and on their attitudes toward gender roles and gender identity. We found no quantitative evidence for any short-term change in teachers self-reported practices as a result of the program, or of any additional impact for the text-message component. Qualitative research showed that attitudes supportive of gender equality did not always align with the traditional ideas of gender roles in the larger community, which was a challenge to gaining broader acceptance of the concepts. These findings suggest that, while teacher training can influence knowledge and attitudes toward gender equality, traditional gender norms can be a barrier to changing behavior in the short term. Such behavioral changes are important in Uganda, where sharp education and gender inequalities persist, particularly in the northeastern region of Karamoja. The primary school completion rate in Uganda is 64 percent, enrollment in lower secondary school is 34.9 percent, and enrollment in upper secondary school is 15.1 percent. Girls are more likely than boys to drop out of school at the higher levels (Pham, Vinck, and Gibbons 2015, 21). Karamoja is overrepresented in the country s lowest development indicators, particularly in education. It has the highest proportion of girls who are not in school or have never been to school, and the highest child mortality and poverty rates 75 percent of households in the region live below the official poverty line (Ministry of Gender, Labour, and Social Development and UNICEF 2015). Statistics show that average years of schooling in Karamoja is as low as three, that there is a 37:1 pupil-teacher ratio, and a 108:1 pupil-classroom ratio (Pham et al. 2015, 21). The context of our study, therefore, is one in which improvements in gender equality are urgently needed. Through our rigorous experimental mixed-methods research, we contribute to the knowledge on what works to improve gender equality in education and peacebuilding, with special reference to conflict-affected settings. To the best of our knowledge, this is the first cluster-rct study of a program for gender socialization in schools within a conflict-affected setting, one that investigates whether gender-sensitive approaches can be introduced into teacher training to reduce gender stereotypes, improve gender equity, and promote peacebuilding. It is also one of the first RCTs that uses a mixed-methods design to determine the effects of development programs in such a setting. 46 Journal on Education in Emergencies

47 CAN TEACHER-TRAINING PROGRAMS INFLUENCE GENDER NORMS? CONTEXT Girls and women in Karamoja endure a hostile environment, due in part to the 20-year conflict between the rebel Lord s Resistance Army (LRA) and the Ugandan government, which lasted until 2006, when the LRA s power declined and peace talks began. During this conflict, tens of thousands of youth were abducted and forced to serve as soldiers, and many girls were victims of forced marriage to combatants, which resulted in deep social and psychological trauma. Children who were abducted often did not return to school, their wealth status later in life was lower, and they reported more symptoms of emotional distress than peers of the same age and location who were not abducted (Annan, Blattman, Mazurana, and Carlson 2011, 889; Blattman and Annan 2010, 882; Opinia and Bubenzer 2011, 5). In addition to this national conflict, Karamoja has a history of recurring conflict between ethnic groups. The region includes seven districts inhabited by at least ten different ethnic groups. Their conflicts are the result of internal economic and social tensions, which often revolve around livestock, particularly cattle. Cattle ownership is a determinant of both social and economic status (Anderson and Broch-Due 1999), and cattle raiding has therefore been prevalent in Karamojong communities (Mkutu 2008). The conflicts are also the result of the region s predominantly pastoral lifestyle. During the dry season, communities tend to migrate to neighboring districts in search of pastures and water for their livestock, which sometimes escalates into border conflicts between tribes that are exacerbated by the different groups proximity to one another. Land disputes are also one of the most common sources of conflict in the region and one of the most difficult to resolve. A 2010 study conducted in four districts in northern Uganda with a representative sample of adults in those four districts reported that 63 percent of all disputes were related to land, and only 48 percent had been resolved by the time of the survey, compared to more than 75 percent of other disputes (Pham and Vinck 2010, 28). Traditional views on gender roles in Karamojong society help to perpetuate violence. Faced with a lack of resources and influenced by traditional ideas about the male breadwinner, men often feel pressured to demonstrate their masculinity by raiding cattle, which fuels violent fighting between clans (Instituto da Defesa Nacional 2013). Men s failure to fulfill their traditional role as provider has also led to psychosocial problems and tension between men and women. These tensions can result in alcoholism, violence against intimate partners, and increased violence July

48 CHINEN, COOMBES, DE HOOP, CASTRO-ZARZUR, AND ELMESKI against women in general (Specht 2013). In addition, the high number of cattle required to win a bride encourages cattle raiding (Vaughn and Stewart 2011). There is a disjuncture between Uganda s formal education system and Karamojong norms. Some Karamojong simply reject formal education (Saminsky 2010). They believe that sending boys to school prevents them from gaining intimate knowledge of their herds, which is where most boys will earn their livelihoods. Girls are similarly expected to perform housework, which has little perceived correlation with what is taught in the classroom. In circumstances where parents must decide whether to send their sons or daughters to school, the girls are often left at home to learn domestic work, marry, and have children. LITERATURE REVIEW Studies of conflict and peacebuilding theorize that education can contribute to gender equality (Winthrop and Kirk 2008). School lesson plans and classroom conduct may play a key role in the transmission or elimination of discrimination (Duncan 2004, 21). For example, teacher training may promote peace by discouraging hostility, and the curriculum can provide positive models of masculinity and femininity that prevent the exacerbation of inequality (Knutzen and Smith 2012). However, education also can undermine gender equality through several mechanisms. Aikman and Unterhalter (2005), for example, state that a lack of adequate toilet facilities and running water can create a barrier to attending school, especially for girls during menstruation. They report further that schools with a limited number of female teachers can create a barrier to girls attendance, as the presence of female teachers tends to be associated with more girl-friendly environments. The nature of teachers pedagogy and teacher-pupil relations can also play a critical role in promoting gender-equitable attitudes and behaviors in their students. For example, teachers and schools can transmit negative gender stereotypes by giving boys more attention in the classroom. Teachers also can undermine the learning experience by using biased language in the classroom, which reinforces gender differences and inequalities. National curricula and textbooks sometimes promote gender stereotypes that lead to gender inequality. Textbooks with stereotypical images of men and women for example, with women depicted as mothers and housewives while men are portrayed in adventurous and influential roles are still common in many countries (Aikman and Unterhalter 2005, 42). Thus, the challenge is not only to change the curriculum content to be more gender sensitive 48 Journal on Education in Emergencies

49 CAN TEACHER-TRAINING PROGRAMS INFLUENCE GENDER NORMS? but also to improve teacher training so teachers are adequately prepared to deliver it (Aikman and Unterhalter 2005, 42). The literature indicates that a first step toward gender equity and peacebuilding should be to identify and transform the widely held norms underlying gender identities and the relations between women and men that reinforce damaging gender and sexual stereotypes (Strickland and Duvvury 2003, 7). However, the peacebuilding community is uncertain how to design programs that address discriminatory gender norms and practices that disadvantage women (Strickland and Duvvury 2003, 8). The rigorous evidence base on which education programs are effective in conflictaffected settings is very limited, particularly on those that reduce gender inequality. Most claims about the relationship between education and gender equality are based on correlational studies whose designs do not allow for addressing counterfactual questions. A systematic review of the literature on education in crisis-affected contexts identified five cluster-rcts and eight quasi-experimental impact studies, none of which focused on outcome measures associated with gender equality (Burde et al. 2015). For example, an RCT in Afghanistan found strong evidence that introducing village schools in Afghanistan resulted in a 52-percentage-point increase in girls enrollment in education and an increase in average test scores of 0.65 standard deviations (Burde and Linden 2013). Another study in Uganda found evidence that community monitoring of the education service provision increased teacher and pupil attendance (Barr, Mugisha, Serneels, and Zeitlin 2012), while a study in Burkina Faso showed that providing schools with girl-friendly water, sanitation, and hygiene facilities increased enrollment by 13 percentage points (Kazianga, Levy, Linden, and Sloan 2013). Most impact evaluations of education programs in conflict-affected settings rely on quantitative designs, and few triangulate the results with qualitative methods (Burde et al. 2015). It thus remains unclear how and why effective programs influence education outcomes. This constraint limits the lessons to be learned from impact evaluations. The lack of mixed-methods research is problematic, because education programs in conflict-affected settings often focus on intangible results that are hard to measure using quantitative research alone (Puri, Aladysheva, Iversen, Ghorpade, and Brück 2015). July

50 CHINEN, COOMBES, DE HOOP, CASTRO-ZARZUR, AND ELMESKI THE PROGRAM The Gender Socialization in Schools pilot program was developed by UNICEF and the Ugandan Ministry of Education, Science, Technology and Sports (MoESTS), which implemented it in partnership with Development Research and Training and the Forum for African Women Educationalists. The pilot was part of UNICEF s Learning for Peace program, which is founded on the idea that education and other social services have strong potential to foster social cohesion and enhance human security in countries affected by and emerging from violent conflict. The training was organized in two stages. First, the implementing partners provided a three-day training of trainers in March 2015 for the coordinating center tutors (CCTs), district school inspectors, and MoESTS personnel. The training explained theoretical concepts of gender, conflict, and peacebuilding using participatory approaches, such as role playing, discussions, and storytelling, that incorporated familiar examples from Karamoja. Second, trained CCTs and inspectors delivered a three-day training for one thousand teachers at central locations in five districts of Karamoja. 1 The training aimed to empower primary teachers as agents of change, promote positive models of masculinity and femininity, and redress gender biases and question social norms. Moreover, the training aimed to create awareness of alternative norms and practices related to gender equality, build skills to help engage pupils in constructive dialogue, and provide materials to foster a shift in gender attitudes and beliefs and promote gender-sensitive practices in the classroom (Development Research and Training 2015). In August/September and November 2015, teachers received refresher trainings to reinforce content. Between April and November 2015, a subset of 276 trained teachers received 13 text messages from UNICEF via the SMS platform GenderTrac. Each text message contained reinforcing reminders for teachers about certain content covered in the training and provided examples of good practices, such as promoting an equitable school environment, mechanisms for conflict resolution in school, positive discipline, and gender-responsive leadership and management. 1 The training modules covered topics such as key gender concepts (e.g., the definitions of gender, sex, gender socialization, gender identities, gender roles, gender equity); gender-responsive teaching approaches (e.g., gender-responsive classroom set-up, gender-responsive language used in classrooms, gender-responsive content delivered by the teacher); gender-responsive learning materials; and gender-responsive classroom interaction. 50 Journal on Education in Emergencies

51 CAN TEACHER-TRAINING PROGRAMS INFLUENCE GENDER NORMS? CONCEPTUAL FRAMEWORK We defined our hypotheses and research questions based on a theory of change that we co-constructed with UNICEF, which maps out the causal chain among activities, outputs, outcomes, and impact, as well as the initial contextual conditions and the assumptions underlying the theory of change (White 2009). The theory of change hypothesizes that generating training materials, conducting teacher trainings, and sending reinforcing text messages about the importance of gender-sensitive education would result in greater knowledge about the difference between gender and sex, and about the relationship between gender, identity, and conflict, which could in turn improve teachers attitudes toward issues of gender equality. These changes could increase the likelihood that teachers would use gender-responsive and peaceful practices. It is important to note that the program focuses on both men and women, which makes the program a relational rather than a single-sex exercise. We hypothesized that this approach would make the program more sustainable, for three reasons. First, including gendered attitudes toward men and masculinity can mitigate male alienation and backlash (e.g., Barker and Schulte 2010; de Hoop, van Kempen, Linssen, and van Eerdewijk 2014; Dworkin, Dunbar, Krishnan, Hatcher, and Sawires 2011). Second, the program s emphasis on positive models of masculinity (as well as femininity) could help to engage men as partners in women s empowerment trajectories. And third, because men are also disadvantaged by norms of negative masculinity (e.g., expectations of participation in raids, fighting, violence), focusing on positive models of masculinity may benefit men and make the communities they live in more peaceful. Several assumptions underlie the theory of change. First, gender equality is a key principle of building sustainable peace. Second, limited socioeconomic and political progress constrain positive shifts in gender norms in conflict-affected areas. Third, education systems offer an institutional platform for instilling gender-equitable ideas and exposing children to positive gender norms. Fourth, teachers, who themselves may be affected by gender bias or perpetuate it, have the capacity to become agents of change by promoting positive visions of masculinity and femininity. The theory of change for the Gender Socialization in Schools program is depicted in Figure 1. July

52 CHINEN, COOMBES, DE HOOP, CASTRO-ZARZUR, AND ELMESKI Figure 1: Theory of Change for the Gender Socialization in Schools Pilot Program Initial Conditions Program Activities Outcomes Impact Weak resilience, social cohesion, and human security in conflict-affected contexts. Conservative gender norms. Lack of discussion among teachers regarding positive masculine and feminine ideals. Women constitute largest proportion of out-of-school youth in Uganda. Assumptions Gender equality is a key principle in peacebuilding. Conflict-affected areas often face constraints that make it difficult to shift to positive gender norms. Education systems offer the largest institutional platform for instilling more genderequitable ideas. Teachers can promote gender equality and interrupt the cycle of violence. Training of trainers: CCT coordinators are trained by UNICEF on gender, conflict, and identity. Materials are produced and distributed to training locations. Training-Only Group (T1): Teachers receive a 3-day training in positive gender socialization and peacebuilding, two refresher trainings, and a teacher handbook. Training + Reinforcing Text Messages Group (T2): Teachers receive a 3-day training in positive gender socialization and peacebuilding, two refresher trainings, and a teacher handbook. Teachers receive biweekly reinforcing text messages. Teachers participate in training. Teachers in the Training + Reinforcing Text Messages Group receive the text messages. Teachers increase knowledge of the information provided in the training. Teachers improve attitudes about gender roles. Teachers improve attitudes toward gender identity and the gender equity. Teacher Practices Teachers promote genderresponsive and peaceful practices. Teachers promote gender equality practices. METHODS Research Design The evaluation employed a cluster-rct in which schools were randomly assigned to the teacher training alone (Treatment 1), the teacher training and reinforcing text messages (Treatment 2), and the control group. We stratified the randomization by the CCTs catchment areas, thus ensuring that each tutor s area included each treatment and control condition. Increasing the geographic proximity of the schools assigned to the study groups was important in accounting for key social norms related to the outcomes of interest and to increase comparability across the groups. To minimize spillovers and contamination, the implementers encouraged local education authorities to minimize information-sharing with the control schools. 52 Journal on Education in Emergencies

53 CAN TEACHER-TRAINING PROGRAMS INFLUENCE GENDER NORMS? The sample was limited to 105 government schools in three of the seven districts in Karamoja, which were evenly assigned to each of the three groups (35 schools in each group). 2 All teachers in the selected schools were invited to participate in the study. This randomization process led to 304, 299, and 313 teachers, respectively, from whom we collected baseline data. Power calculations that took into consideration the nested structure of the evaluation suggest that the study had 80 percent power to detect a minimum effect size of 0.28 standard deviations. 3 Figure 2 presents the flow diagram of the RCT design. Figure 2: Flow Diagram of the Randomized Controlled Trial Design Government Schools in Three Districts of Karamoja: 105 Schools Randomized Baseline Training-Only (T1) Group: 35 Schools 299 Teachers Training + Reinforcing Text Messages (T2) Group: 35 Schools 304 Teachers Control Group: 35 Schools 313 Teachers Endline 35 Schools 201 Teachers 35 Schools 224 Teachers 35 Schools 225 Teachers 2 These three districts were selected as follows. First, two of the seven districts were excluded by implementers because they were already benefiting from several other education programs. Second, to minimize the length and cost of data collection, the research sample was selected from the three districts where training was scheduled to happen first. Within the CCTs catchment areas in each of the three districts, we first randomly sampled the maximum number of schools that was a multiple of three, and then we randomly assigned those schools to each of the three study groups to have a balanced design. 3 The intra-class correlation for the different outcome measures was approximately We conducted all power calculations using Optimal Design software. July

54 CHINEN, COOMBES, DE HOOP, CASTRO-ZARZUR, AND ELMESKI Quantitative Teacher Survey We developed the survey based on a comprehensive literature review and a review of best practices in measuring outcomes depicted in the theory of change. The items are consistent with other inventories used to measure gender attitudes. 4 The instrument was then refined on the basis of several consultations with UNICEF, MoESTS staff, and key stakeholders, followed by three pilots of the instrument in the field. During these pilots, we paid close attention to the length of the survey, the respondents understanding of the survey questions, and some basic psychometric properties of the instruments. These pilots enabled us to create an instrument informed by the specific contextual characteristics of Karamoja. During this time, we also added vignettes to the questionnaire that we developed in collaboration with local Karamojong people (e.g., teachers, education officials, local data collectors). We did not rely on previously validated tests because they were not available for the context of Karamoja or the specific goals of the program. The final survey included questions related to gender norms, the division of household and labor duties between men and women, and differences in boys and girls educational opportunities and experiences. We used vignettes to measure teachers attitudes toward gender norms on topics such as gender roles in the household and sexual and physical violence. These vignettes described a fictional scenario and were typically used to determine the ways people make judgments and decisions about sensitive topics. Using vignettes can reduce the likelihood of courtesy and social desirability bias (White and Phillips 2012). Finally, survey items were worded to be consistent with Ugandan proverbs, folklore, and literature on the roles of men and women in Karamoja. In the final version of the questionnaire, we used two groups of questions to measure teachers knowledge of the information provided in the training. The first questions captured whether teachers understood the difference between gender a social construction regarding the roles of women, girls, men, and boys and sex, which refers to the biological characteristics of being female or male. The second group of questions focused on knowledge about other topics covered in the training, including gender-sensitive lesson planning, the legal framework for equal access to education, and the relationships among gender equality, peacebuilding, and social cohesion. 4 Examples of these inventories include Ashmore, Del Boca, and Bilder (1995); Glick and Fiske (1997); Baber and Tucker (2006); and Pulerwitz and Barker (2008). 54 Journal on Education in Emergencies

55 CAN TEACHER-TRAINING PROGRAMS INFLUENCE GENDER NORMS? We created several index scores to measure teachers attitudes toward gender roles, gender identity, gender equality in schools, and several dummy indicators to measure sexual harassment of girls by boys, violence by boys against girls, and violence by girls against boys. The gender-role indexes measured the shared expectations of behavior based on a person s gender, while the gender-identity index aimed to capture how masculine and feminine teachers see themselves in terms of what it means to be a man or a woman in their society. The items in the gender-equality construct captured attitudes toward gender equality in school and whether teachers expectations for girls and boys were similar. The binary variables measured whether teachers punished boys for sexually harassing girls, whether teachers punished boys or girls for behaving violently toward the opposite sex, and whether teachers punished boys and girls equally for behaving violently toward the opposite sex. Finally, we generated two indexes to approximate teachers practices. The first index measured teachers gender responsiveness when planning and implementing activities and exercising discipline in the school, while the second measured teachers practices associated with gender equality (e.g., we asked teachers whether they assigned more difficult tasks to boys or easier tasks to girls, and whether they used the same strategies to teach girls and boys). The survey also included questions related to the school s cultural practices, such as relationships between teachers and students, relationships between boys and girls, and school clubs. We also included several questions to capture teachers demographic and teaching backgrounds and basic school characteristics, such as the number of teachers and students, and information on infrastructure and vailable services. Table 1 summarizes the contents of the teacher questionnaire. 5 5 To construct the outcomes of interest, we generated an index and a scale. Correct knowledge answers or progressive attitudes were coded as 1, and the index was computed as the summation of these answers. Practices were coded using a scale of 1 to 4, where responsive and peaceful practices received higher scores. The scale was created through a factor analysis in which we constructed weights from the matrix of factor loadings. The Cronbach s alpha coefficients, which measured the internal consistency of the outcome indexes, ranged from 0.62 to 0.73 for the knowledge indexes, from 0.64 to 0.88 for the attitudes indexes, and from 0.68 to 0.72 for the practice indexes. Finally, for the descriptive indexes, the Cronbach s alpha coefficients ranged from 0.71 to Evidence suggests that Cronbach s alpha coefficients smaller than 0.7 may be an indication of less reliable scales. We did some robustness checks in which we conducted the same impact analysis for individual items. These analyses suggest that the results of the study are robust to the use of individual items as opposed to indices or scales. These results are included in the report created for UNICEF (Chinen et al. 2016). In this paper, we present the program s impact on the index because the interpretation of these results is more intuitive. Nonetheless, the results are robust to the use of scales as outcome measures. July

56 CHINEN, COOMBES, DE HOOP, CASTRO-ZARZUR, AND ELMESKI Table 1: List of Outcome Indexes and Indicators Created from the Teacher Survey Outcome Indexes 1 Knowledge about the difference between gender and sex 2 Knowledge about gender, identity, and conflict 3 Attitudes toward gender roles, Index 1 4 Attitudes toward gender roles, Index 2 5 Attitudes toward gender roles, Index 3 Description Teacher understands the difference between gender and sex. Teacher understands issues of gender, identity, and conflict. Teacher believes men and women are equally capable of conducting jobs that are traditionally associated with one gender. Teacher supports textbook image that shows a father in a caretaker role. Teacher does not oppose being seen conducting activities traditionally associated with women. 6 Attitudes toward gender identity Teacher disagrees with statements describing very traditional masculine stereotypes. 7 Attitudes toward gender equality Teacher agrees with statements associated with gender equality. 8 Gender-responsive and peaceful self-reported practices 9 Gender equality self-reported practices Teacher conducts gender-responsive and peaceful practices in the classroom, when planning, implementing activities, managing behavior, and exercising discipline in the school. Teacher conducts activities to promote gender equality in the classroom. Additional Dummy Indicators 10 Attitudes: Reactions to sexual harassment (vignettes) 11 Attitudes: Fair punishment to sexual harassment (vignettes) 12 Attitudes toward violence (vignettes) Teacher intervenes in scenarios of sexual harassment. Teacher blames harasser for harassment. Teacher punishes harasser. Teacher opposes violent retaliation to sexual harassment. Teacher punishes males and females correctly and equally. Teacher intervenes in scenarios of classroom violence. Secondary and Long-Term Indexes 13 Gender and culture in schools Teacher identifies positive gender culture in the school. 14 Problems in the school environment Teacher identifies many gender-based problems in the school environment. 15 Teachers sense of self-efficacy Teacher feels capable of solving gender-based problems in the school environment. 56 Journal on Education in Emergencies

57 CAN TEACHER-TRAINING PROGRAMS INFLUENCE GENDER NORMS? Data collection in the treatment group took place in the teacher-training locations on the first day, before the teachers started to receive the training. All teachers who attended the training agreed to complete the survey. The data collectors organized the teachers in classrooms, explained the goals of the study, read the consent agreement aloud, and stayed in the classroom to answer questions or clarify aspects of the survey. Data collection in the control group took place in schools from Monday to Thursday during the same weeks the intervention group data was collected. All teachers in the control schools agreed to participate in the survey. As with the treatment group, the data collectors gathered the teachers in classrooms and followed the same protocols, thus standardizing the datacollection process for the two groups to the extent possible. 6 Qualitative Instruments We collected qualitative data from the CCTs who were present during the training and from head teachers whom the implementing partners identified as teacher leaders, based on their participation in the training and their interactions with other participants. We conducted one-hour semi-structured interviews at midline data collection with 15 intervention CCTs and 8 head teachers. The purpose of these interviews was to understand leaders experiences with (a) implementation of the training, (b) school-level follow-up on the training, and (c) the level of understanding of training concepts. In order to facilitate a deeper understanding of the factors that enabled, or inhibited, the uptake of gender-equitable practices and peaceful conflict resolution as a result of this program, we collected qualitative data from teachers and students in a random selection of six intervention and two control schools across the three districts and the three study groups. Endline focus group discussions (FGDs) with the treatment teachers built on midline discussions, primarily about the training itself and about the specific challenges the teachers had experienced in applying practices. The endline FGDs specifically targeted knowledge, attitudes, and practices that might have changed as a result of the program. The FGDs with two control schools allowed us to find contrasts between treatment teachers and control teachers. The interviewers facilitated the approximately two-hour FGDs using a guide, which included a flexible set of questions and probes intended to invite the participants to steer the discussion toward the issues that interested them while ensuring that they remained focused on relevant topics. We organized separate focus groups for male and female teachers so participants would feel 6 The study team received consent from all adult participants and handled the quantitative and qualitative data according to procedures and protocols approved by the institutional review board. July

58 CHINEN, COOMBES, DE HOOP, CASTRO-ZARZUR, AND ELMESKI free to express their authentic experiences. We targeted six participants in each focus group, although some groups were smaller because of the low number of female teachers. Analyses We examined the main effects the interventions had on the outcomes of interest using an analysis of covariance (ANCOVA) model. The ANCOVA model allows for an estimate of the causal effect of the program by comparing outcomes in the intervention schools with outcomes in the control schools, and controlling for the value of the outcome variable at baseline. The main advantage of the ANCOVA model, as opposed to difference-in-difference analysis, is that the ANCOVA model increases statistical power, particularly when outcomes are not strongly autocorrelated (McKenzie 2012, 211). The model was particularly appropriate for this study because we changed a few items and the wording of a few questions between the baseline and the endline, based on feedback and analyses of the baseline data. Using an ANCOVA model enabled us to use the original index as control variables, despite the changes in the wording. This would not have been possible using a difference-in-difference model. The ANCOVA model used cluster robust standard errors at the school level to account for the nested structure of the data. We checked the robustness of the treatment estimates against several different model specifications that included a different set of covariates, and found that the impact estimates were robust to the specification of the regression. 7 The impact results for the statistical model that only controlled for the pretest score are presented in this paper. We also investigated possible selection bias due to missing data, because 29 percent of the teachers who participated in the baseline were not available to complete the endline survey. For this reason, we examined whether the percentage of teachers with missing data was similar or different across the three study groups; whether teachers with complete baseline and endline data were equivalent in terms of observed covariates collected at baseline across the three study groups; and whether teachers with missing data were similar at baseline to those with complete data. These analyses revealed that the percentage of teachers with missing data were similar across the three groups, and that teachers with completed data at both baseline and endline were equivalent in the vast majority of 7 We specified a total of six model specifications for each outcome variable. The first model included only the treatment indicator; the second one added the pretest score; the third one added the district fixed effects; the fourth included teacher characteristics such as gender, religion, and education; the fifth model added school characteristics, such as the number of female teachers and enrollment size; and the sixth combined treatments 1 and 2, in light of the lack of differential treatment effects across the two interventions. 58 Journal on Education in Emergencies

59 CAN TEACHER-TRAINING PROGRAMS INFLUENCE GENDER NORMS? observed characteristics gathered in the teacher survey. 8 We also investigated the possibility of spillovers (or the possibility that the interventions affected control teachers), and found limited evidence of spillover effects. 9 Finally, we explored the possibility of heterogeneous treatment effects by teachers sex. The robustness analyses and complete impact regression results are presented in the endline report we developed for UNICEF (Chinen et al. 2016). To analyze the qualitative data, we used methodological triangulation and triangulation among raters to ensure data trustworthiness and credibility of findings. The research team utilized grounded theory (Glaser and Strauss 2009) to guide analysts trained in NVivo qualitative data analysis Software (QSR International Pty Ltd., Version 10, 2012). Grounded theory utilizes qualitative data to deduct a new theory about the findings, rather than testing an existing one. Combining the grounded theory with a rigorous impact-evaluation design enabled us to triangulate the research findings through mixed methods. The quantitative research served to test predefined hypotheses, while the qualitative research enabled us to gain a better understanding of why the program positively influenced some, but not all, outcomes of interest. Three raters separately coded the text data to independently identify the themes in the discussion. These themes formed the coding structure used to categorize raw data from the interviews and focus groups and identify themes about the primary findings. An inter-rater reliability test showed that the three primary coders had an overall average of 99.1 percent agreement, indicating a high level of consistency among the researchers in interpreting the data and clarity of the coding scheme. 8 We called absent teachers and inquired why they missed the endline survey. The three main reasons were the following: (1) teachers were not informed about the teacher training or came late to training (in the case of intervention teachers); (2) teachers were sick during the data collection; and (3) teachers were busy grading primary leaving exams. A list of all the reasons and all the missing analyses are presented in the endline report (Chinen et al. 2016). 9 Approximately 33 percent of control teachers (or 76 teachers) indicated receiving some training in gender, conflict, and peacebuilding, and 20 percent reported receiving another training in gender issues. Moreover, 18 percent of control teachers reported receiving some coaching on gender, conflict, and peacebuilding, and 7 percent indicated receiving some text messages about gender, conflict, and peacebuilding. These results suggested the presence of spillovers from treatment to control teachers, which may have resulted in an underestimation of the impact of the intervention. However, we did not encounter additional evidence that control teachers attended the trainings. Besides, it is possible that some teachers confused the different gender trainings. Approximately 49 percent of the 76 control teachers who reported attending the Gender, Conflict, and Peacebuilding training also reported attending another training in gender issues. July

60 CHINEN, COOMBES, DE HOOP, CASTRO-ZARZUR, AND ELMESKI RESULTS Balance at Baseline At baseline, we found that the treatment and control groups were statistically equivalent in the vast majority of observable characteristics. In other words, the randomization process successfully created equivalent groups before the intervention started. We found no statistically significant differences between treatment and control teachers. Tables 2 and 3 present descriptive statistics for teachers demographic and background characteristics, as well as teachers education characteristics. The descriptive statistics for all the survey covariates, including for the baseline outcomes, can be found in the Baseline Report generated for the study (Chinen et al. 2016). Table 2: Baseline Teacher Demographics and Background Characteristics Variable Control Training-Only (T1) Training + Text (T2) p-values Diff (SD) N Mean N Mean N Mean C = T1 C = T2 T1 C T2 C (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) Gender: Female Ethnicity: Ethur Other Bokora Karimojong Religion: Catholic Protestant Other Native language: Ngakaramojong Thur Ateso Luo Marital status: Officially married Unofficially married Other main source of income: None Growing crops Journal on Education in Emergencies

61 CAN TEACHER-TRAINING PROGRAMS INFLUENCE GENDER NORMS? (Table 2 cont.) Control Training-Only (T1) Training + Text (T2) p-values Diff (SD) N Mean N Mean N Mean C = T1 C = T2 T1 C T2 C (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) Teacher s age Years teacher lived in the district where born Number of women married to Total number of children Female children Male children NOTES: Robust t-statistics clustered at the school level. Table 3: Baseline Teacher Educational Characteristics Variable Control Training-Only (T1) Training + Text (T2) p-values Diff (SD) N Mean N Mean N Mean C = T1 C = T2 T1 C T2 C (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) Teaching language: English Swahili Local language Opinion matters in the community where teaching Meet with pupils parents: At least once a month Once per term Received any previous training on gender issues Years teaching Years teaching in that school Total grades taught Minutes to go from home to school NOTES: Robust t-statistics clustered at the school level. July

62 CHINEN, COOMBES, DE HOOP, CASTRO-ZARZUR, AND ELMESKI Impact on Knowledge Quantitative evidence indicated that the program resulted in an increase in teachers knowledge of the information provided in the training. We found evidence for a positive and statistically significant program impact on teachers knowledge about the difference between gender and sex. The point estimates were 0.35 (p < 0.1) and 0.60 (p < 0.01) for the training-only and training-plus-texting groups, respectively. This finding indicates that the effect for the training-plustexting group was almost twice the effect for the training-only group, but this difference was not statistically significant (p < 0.22). We found no positive effects on the second indicator of knowledge about gender, identity, and conflict. Table 4 presents the impact on knowledge. Qualitative evidence indicated similarly that, at endline, teachers were better able to use the concepts of gender and sex consistently with the training program s definitions (e.g., gender sensitive, gender socialization). Teachers also understood that classroom discrimination based on gender identity could affect social interactions, girls self-confidence, and their feeling that they need to skip school during menstruation. In describing the need to address the stigma of menstruation, one teacher said, I even talk to them about fear; they should not have fear when they are undergoing their menstruation period. Finally, teachers identified ways to ensure a more gender-sensitive environment, including the classroom set-up, equal participation and representation in lessons, and shared responsibility. However, qualitative data also suggested that challenges remained. One participant pointed to the difficulty of learning the concepts: There were certain concepts that were... difficult to define or explain; [for example], when it came to things like gender disparity, gender equality, and all those concepts. The same participant also pointed out the value of refresher trainings in helping to reinforce concepts and ensure understanding of basic concepts: When we first started with the first training, it was like [we] were trying to understand those concepts about gender, but the second one was so much [more] interesting and most participants [expressed interest in having] another time of really going through those concepts again. 62 Journal on Education in Emergencies

63 CAN TEACHER-TRAINING PROGRAMS INFLUENCE GENDER NORMS? Table 4: Impact on Knowledge Outcomes Knowledge Understands the difference between gender and sex Understands issues of gender, identity, and conflict Control Mean (Endline) Program Impact Training-Only (T1) SE Effect Size Training + Reinforcing Text Messages (T2) Program Impact SE Effect Size * *** NOTES: * p < 0.1; ** p < 0.05; *** p < Adjusted regression results using ANCOVA OLS controlling for pretest score. Cluster robust standard errors at the school level. Sample size ranges between 566 and 630, depending on outcome. Impact on Attitudes Overall, the quantitative and qualitative evidence indicated that the program resulted in teachers having more positive attitudes toward gender roles and gender identity. We found positive and statistically significant program effects on all three indicators of teachers attitudes toward gender roles, measuring shared expectations of behavior given one s gender. The results for all attitudes indexes are presented in Tables 5 and 6. The first gender-role index was created by asking teachers whether women or men are more or equally capable of doing jobs that are traditionally associated with one gender. Intervention teachers were more likely to agree with statements that implied relatively progressive attitudes toward gender roles. For example, intervention teachers were more likely to agree with statements suggesting that men and women are equally capable of doing jobs traditionally associated with one gender (e.g., engineer, mechanic, nurse, politician). The ANCOVA point estimates on the full scale (which ranged from 0 to 10 points) were 0.83 (p < 0.01) for the training-only group and 0.48 (p < 0.05) for the training-plus-texting group, both of which were statistically significant. The second gender-role index was created by presenting a hypothetical vignette in the proposed new English textbook. It showed a father cooking dinner and looking after his baby to capture teachers attitudes toward traditional gender roles (see Figure 3). The teachers were then asked whether they would support having such a picture in the textbook, whether this was an example of gender equality that they would promote, and whether they would think this man s wife was treating him well if he were their brother. We found that intervention July

64 CHINEN, COOMBES, DE HOOP, CASTRO-ZARZUR, AND ELMESKI teachers were also more likely to show more progressive attitudes toward gender roles depicted in the hypothetical situation. The ANCOVA point estimate for the index (which ranged from 0 to 3) was 0.44 (p < 0.01) for the training-only and the training-plus-texting groups, which were both positive and statistically significant. Figure 3: Hypothetical Situation Depicted in a Vignette in the Proposed New English Textbook Showing a Father Cooking Dinner and Looking after His Baby The third gender-role index measured teachers attitudes toward conducting activities traditionally associated with women. This index was considered relevant, given that most of the teachers (75 percent) were men. For example, intervention teachers were more likely to disagree with statements such as, I would not want my friends to see me washing women s clothes and I would not want my friends to see my spouse correcting me in public. The pattern of responses for other items was more similar among the three groups. The regression point estimates for the index scale (which ranged from 6 to 24 points) were 0.59 (p < 0.1) for the training-only group and 0.57 (p < 0.01) for the training-plus-texting group, both of which were positive and statistically significant. Positive quantitative effects were also found for teachers attitudes toward gender identity, which refers to how male or female teachers see themselves relative to what it means to be a man or a woman in their society. Intervention teachers were more likely to disagree with very traditional masculine stereotypes, including such statements as, Some women need to be beaten, Educated women make unruly wives, When you beat boys, you raise disciplined men, When men are speaking, serious woman are not supposed to talk. The ANCOVA point estimates for the full index scale (which ranged from 22 to 52 points) were 1.26 (p < 0.01) for the training-only group and 0.62 (p < 0.1) for the training-plus-texting group. 64 Journal on Education in Emergencies

65 CAN TEACHER-TRAINING PROGRAMS INFLUENCE GENDER NORMS? We found limited and no evidence for positive program effects on the index we used to measure attitudes toward gender equality, and on attitudes toward sexual harassment and on punishment for sexual harassment or violence. We used vignettes to understand the action teachers might take in a hypothetical situation about sexual and physical violence, and in a conflict between boys and girls in the classroom. The vignettes included situations in which a boy inappropriately touched a girl, a girl was physically violent, and men were showing behavior usually associated with women. Although the direction of the point estimates was generally positive, the results were generally not statistically significant. Thus, the findings suggested little evidence of changes in reactions to sexual harassment or in the punishment of students for sexual harassment or violence. Table 5: Impact on Attitudes toward Gender Roles, Gender Identity, and Gender Equality Attitudes Gender Roles Index 1: Believes women and men are equally capable of doing traditionally gendered jobs Gender Roles Index 2: Supports textbook image of father in a caretaker role Gender Roles Index 3: Does not oppose being seen conducting activities traditionally associated with women Gender Identity: Disagrees with statements of masculine stereotypes Gender Equality: Agrees with statements of gender equality Control Mean (Endline) Training-Only (T1) Program Impact SE Effect Size Training + Reinforcing Text Messages (T2) Program SE Effect Size Impact *** ** *** *** * ** *** * * NOTES: * p < 0.1; ** p < 0.05; *** p < Adjusted regression results using ANCOVA OLS controlling for pretest score. Cluster robust standard errors at the school level. Sample size ranges between 548 and 613, depending on outcome. July

66 CHINEN, COOMBES, DE HOOP, CASTRO-ZARZUR, AND ELMESKI Table 6: Impact on Attitudes toward Sexual Harassment, Punishment for Sexual Harassment or Violence Attitudes: Additional Dummy Indicators Reactions to Sexual Harassment: Intervenes in scenarios of sexual harassment Control Mean (Endline) Program Impact Training-Only (T1) SE Effect Size Training + Reinforcing Text Messages (T2) Program Impact SE Effect Size Blames harasser for harassment Punishes harasser ** Opposes violent retaliation to sexual harassment Fair Punishment to Sexual Harassment: Punishes females and males correctly and equally Attitudes Toward Violence: Intervenes in scenarios of classroom violence ** ** NOTES: * p < 0.1; ** p < 0.05; *** p < Adjusted regression results using ANCOVA OLS controlling for pretest score. Cluster robust standard errors at the school level. Sample size ranges between 636 and 648, depending on outcome. The qualitative data yielded similarly mixed results on attitudes. Teachers basic attitudes changed, as they reported that boys and girls should be equal in responsibilities, work, and their futures. The majority of teachers said children should not be encouraged to participate in girl-only or boy-only activities and should share responsibilities. Several teachers observed that the trainings expanded their ideas about what girls could do in the classroom. One teacher said, I went through a girl s school throughout my education... we used to say that boys do more work than girls, but with training and these techniques, we have come to learn that all these people are equal. Several also noted that, after encouraging girls in math, they saw the girls performance improve sometimes beyond that of the boys. However, some of the teachers indicated that traditional attitudes continue to shape some of their approaches in the classroom. For example, one head teacher said, Of course you know culture is part of the community, [and] as far as our community is concerned... there are certain responsibilities which are more of men than of ladies. 66 Journal on Education in Emergencies

67 CAN TEACHER-TRAINING PROGRAMS INFLUENCE GENDER NORMS? In sum, the program meaningfully changed teachers attitudes toward gender roles, but these changes also created challenges for them, given an environment in which traditional gender norms heavily dictate children s roles and responsibilities. Results on Self-Reported Practices The program did not appear to influence overall teacher practices in the short term. No consistent quantitative evidence was found for positive effects on the two overall indexes of self-reported practices that were created from the survey. The intervention and control teachers gave similar answers to most of the questions included in the two indexes. These results are presented in Table 7. The first index of self-reported practices, Gender-Responsive and Peaceful Practices, measured teachers gender responsiveness when implementing activities, managing behavior, and disciplining boys and girls separately. For example, we asked teachers whether they rewarded girls for behaving appropriately, whether they ensured that girls used peaceful means to resolve conflicts with their peers, how teachers disciplined girls who misbehaved, whether teachers helped girls catch up on the lessons when they were absent for genuine reasons, etc. The same questions were also asked to inquire about their practices with boys. For the Gender-Responsive and Peaceful Practices index (which ranged from 12 to 48 points), the ANCOVA point estimate for the training-only group was -0.12, but it was not statistically significant. The point estimate for the training-plus-texting group was 0.57 (p < 0.1), which was statistically significant at the 10 percent level. The second index of self-reported practices, Gender-Equality Practices, measured teachers practices related to gender equality. For example, we asked teachers whether they assigned more difficult tasks to boys or easier tasks to girls, whether they discussed strategies for providing a safe learning environment for girls and boys with other teachers, whether they used the same strategies to teach girls and boys, whether they made sure girls and boys had equal opportunities to participate, etc. For the Gender-Equality Practices index (which ranged from 15 to 60 points), the ANCOVA point estimates were 0.07 for the training-only group and 0.39 for the training-plus-texting group. These effects were not statistically significant. July

68 CHINEN, COOMBES, DE HOOP, CASTRO-ZARZUR, AND ELMESKI Table 7: Impact on Practice Outcomes Practices Conducts gender-responsive and peaceful practices in the classroom Conducts activities to promote gender equality in the classroom Control Mean (Endline) Program Impact Training-Only (T1) SE Effect Size Training + Reinforcing Text Messages (T2) Program Impact SE Effect Size * NOTES: * p < 0.1; ** p < 0.05; *** p < Adjusted regression results using ANCOVA OLS controlling for pretest score. Cluster robust standard errors at the school level. Sample size ranges between 567 and 587, depending on outcome. Qualitative data indicated that intervention teachers adopted some practices taught in the training. The teachers were best able to grasp more pragmatic training concepts, such as equitable classroom set-up, which mainly involved a mixed-gender seating arrangement. One teacher said, If you have 10 boys and 10 girls, they should sit in [an] arrangement whereby a boy and a girl sit [in] the reading corners. Some teachers reported fostering equal representation by dividing resources (such as textbooks) equitably between boys and girls, and encouraging equitable participation in class activities. Other teachers reported dividing classroom responsibilities between boys and girls, including leadership roles and classroom duties. Several teachers explained that gender-sensitive lesson planning meant creating lessons that had objectives, activities, and examples that incorporated both boys and girls. One teacher said, The technique that I am now applying in the classroom situation is... considering both boys and girls equally, while another said, Both boys and girls have to participate in the lesson. Some teachers mentioned that they should tailor lessons to boys and girls, saying, for example, I also learnt about friendly methods which can make a child really participate in an activity, and also [about] the instructional materials, which should be child friendly. The activity which is given should cater for all without any gender discrimination. However, most teachers did not explain how they tailored lessons to male and female needs or connected gender-equitable practices to peacebuilding. In addition, data indicated that traditional practices such as corporal punishment are still used in the classroom. The short timeline of the program may have limited its ability to promote more complex changes in ideas and practices. 68 Journal on Education in Emergencies

69 CAN TEACHER-TRAINING PROGRAMS INFLUENCE GENDER NORMS? The teacher training promoted practices by drafting action plans that encourage a peaceful school environment for students. Although teachers seemed to have a basic understanding of the purpose and use of action plans, not all were aware of how each element was linked to gender. They also expressed difficulty meeting the goals of the plans because of larger issues related to poverty in their schools. One teacher said, for example, I talked of the [seating] arrangement in our action plan, but the challenge we have faced with this is inadequate [seating] facilities. Results on the Effect of Complementary Text Messages We did not find consistent evidence across the different outcome measures that teachers who received reinforcing text messages in addition to the training activities earned more positive scores than teachers who only received the training. These results could be partially explained by the fact that 28 percent of the teachers in this group reported not receiving any text messages. Nonetheless, teachers on average reported receiving 13 messages related to the program, which is the number of reinforcing messages sent by UNICEF. The finding suggests that, at least in the first eight months of implementation in which they received three trainings, the SMS program did not bring additional benefits to the teacher training. Results on Secondary and Long-Term Measures Our results showed no evidence of positive effects from the program on the measures Gender and Culture in Schools, Problems in the School, and Teachers Sense of Self-Efficacy to solve the most pressing problems of the school. The first outcome measured general aspects of the school culture. For example, we asked teachers whether they knew about their pupils families, whether they talked to children about their personal lives, whether they organized clubs, etc. The second outcome inquired about common problems in the school environment. For example, we asked whether hunger, absenteeism, or early marriage were common problems in their school. Finally, the third outcome attempted to measure whether teachers thought they were capable of resolving the problems listed in the previous outcome. We did not expect any statistically significant effects of the program on these outcome measures because the program did not aim to affect them in the short term. Courtesy and social desirability bias might nonetheless have resulted in statistically significant differences between the treatment and control teachers. That we did not find these statistically significant differences suggests that courtesy July

70 CHINEN, COOMBES, DE HOOP, CASTRO-ZARZUR, AND ELMESKI and social desirability bias did not create a high degree of bias in our positive impact estimates on knowledge and attitudes. These results are presented in Table 8. Table 8: Impact on Secondary and Long-Term Outcomes Secondary and Long- Term Measures Identifies positive gender culture in school Identifies many genderbased problems in the school environment Feels capable of solving gender-based problems in the school environment Control Mean (Endline) Program Impact Training-Only (T1) SE Effect Size Training + Reinforcing Text Messages (T2) Program Impact SE Effect Size NOTES: * p < 0.1; ** p < 0.05; *** p < Adjusted regression results using ANCOVA OLS controlling for pretest score. Cluster robust standard errors at the school level. Sample size ranges between 574 and 576, depending on outcome. The qualitative findings indicate that teachers recognized the need for greater support and found it difficult to independently obtain support from parents, politicians, and other community leaders. Although many teachers have made progress in their understanding of gender equality, the majority continue to have difficulty reconciling these concepts with traditional ideas of gender in the larger community. The disjunction between the training ideas and the deeply embedded community norms was evident throughout the data. Multiple teachers expressed difficulty promoting equal opportunity and sharing responsibility in a community where gender norms heavily dictated children s roles and responsibilities. One teacher described the difficulty with involving parents: Problem comes when they disagree with you and I don t know which means we should use, because we can also not force them... so the influence I think is on the ground. Teachers said they should involve parents in gender and peacebuilding because, as one teacher noted, it is from the community that [students] should learn first... before they come to school. In response to the resistance teachers faced in implementing training concepts, some had already involved community members by holding PTA meetings and engaging school management. Although few teachers seemed to have engaged students or the community in sensitization activities, several were able to describe avenues they could take 70 Journal on Education in Emergencies

71 CAN TEACHER-TRAINING PROGRAMS INFLUENCE GENDER NORMS? to sensitize stakeholders on particular issues. Several other teachers mentioned the importance of communicating rules and expectations about conflict in the classroom to the students. Results by Teachers Sex Finally, we did not find evidence for statistically significant differences in the estimates of impact between male and female teachers. This lack of evidence for the heterogeneity of effects could be explained by the small number of female teachers in the schools (25 percent), which could have resulted in a lack of statistical power to detect subgroup effects. 10 DISCUSSION This paper offers evidence that teacher-training programs that emphasize gender socialization can have a positive short-term impact on knowledge about the difference between gender and sex, and on attitudes toward gender roles and gender identity. The study, however, found no quantitative evidence for a short-term positive impact on self-reported teacher practices. The triangulation of qualitative and quantitative findings indicated that, after eight months, the program equipped intervention teachers with new knowledge about gender, changed their basic attitudes about gender equality issues, and taught them about less traditional views on gender roles. However, the more in-depth qualitative data suggested that teachers still identified with traditional gender norms and beliefs about gender. These findings imply that more reinforcement, longer-term programming, or more community involvement is required to encourage teachers to successfully transfer what they learned to real-life situations in school and non-school settings. Traditional ideas of gender in the community can be a barrier to changing complex behavior in the short term. The disjunction between the training ideas and deeply embedded community norms was evident throughout the data. Teachers reported that they had difficulty enforcing new ideas about gender norms that did not align with traditional conceptions of gender. The short timeline of the program may also have limited its ability to promote more complex changes in ideas and practices. The majority of the teachers in our sample continue to have difficulty reconciling views consistent with gender equality with traditional ideas of gender in the larger community. 10 Of the 916 teachers at baseline, 687 were male and 229 were female. July

72 CHINEN, COOMBES, DE HOOP, CASTRO-ZARZUR, AND ELMESKI The lack of evidence for positive effects on self-reported teacher practices is consistent with research suggesting that teacher practices are difficult to change in the short term (Bonder 1992; Mukhopadhyay and Wong 2007; Stromquist 2007; Sullivan 2013; World Bank 2008). However, the qualitative component showed that intervention teachers appeared to be changing some simple classroom practices, such as seating arrangements. These positive changes may have resulted from concrete examples in the training sessions and training manual that explain how to operationalize the training concepts. Although previous research indicates that sending reminder SMS messages can encourage people in developing countries to increase their financial savings (Karlan, McConnell, Mullainathan, and Zinman 2010), we did not find evidence that the complementary text messages sent to reinforce the information communicated during the teacher training had a positive impact. Uganda has been using SMS successfully to improve communication between education stakeholders by communicating exam results from the National Examination Board (Ndiwalana 2011). It is possible, however, that messages about gender equality are too complex to communicate via SMS messages. It seems important to reconsider the content of these messages, how they are delivered, and any limitations on teachers ability to access them. The findings have several implications for policy and practice. Perhaps most importantly, they speak to the importance of community involvement by parents, politicians, and other community leaders in creating a more enabling environment in which new ideas can be welcomed, accepted, and translated into practice. In addition, the prevalence of qualitative data on the challenge of reconciling traditional norms enforced in the home with new ideas presented in school suggests that the more direct involvement of community and school governance bodies could enhance local buy-in for the content of the training. Finally, the training could benefit from providing teachers with regular coaching, on-site monitoring visits, and/or one-on-one reflection sessions. The literature on the effectiveness of teacher training demonstrates that one-time in-service trainings at a central location tend to be less effective than longer-term teacher-training strategies (Conn 2014; McEwan 2001; Showers and Joyce 1996). Such long-term strategies may be even more of an imperative in programs that aim to change social norms, such as the Gender Socialization in Schools pilot program. We need to remain cautious in interpreting these findings because of several limitations. First, the program was evaluated during its first year of implementation. New programs may experience unexpected challenges, or they may not be 72 Journal on Education in Emergencies

73 CAN TEACHER-TRAINING PROGRAMS INFLUENCE GENDER NORMS? implemented as intended, both of which could weaken the effects of the program in the first years. Second, we were only able to estimate the short-term effects of the program (after eight months). Third, limited resources precluded our collecting data from teachers by visiting all the treatment schools. We mitigated this limitation by administering the survey the morning the teachers arrived for the training. This strategy reduced data-collection costs but prevented gathering data in exactly the same way across the three study groups (the research team did attempt to mimic the same conditions and procedures across groups). Fourth, the limited resources and short timeline limited our focus to teachers, who were the program s direct beneficiaries. Finally, quantitative data on the teachers were limited to self-reported surveys. Interviews and other more comprehensive datacollection methods were exclusively qualitative. To overcome the limitations of self-reported data which may suffer from courtesy and social desirability bias we used vignettes, minimized the use of leading questions, and included various types of questions. We also piloted the instrument multiple times and revised some items after considering comprehensive feedback from local experts. Future research should focus on experimental designs to determine the impact of gender-socialization programs at the student level and the longer-term effects on teacher practices. We argue that the impact of these programs may be different in conflict-affected settings because psychosocial development and social cohesion may play a smaller role in other low- and middle-income countries. It therefore will also be important to compare the effectiveness of these programs between conflict-affected countries and other low- and middle-income countries. Such comparisons can provide lessons about the links between conflict, gender equality, and social cohesion. In addition, it will be crucial to improve the measurement of teacher practices related to gender and of peacebuilding indicators so we can more fully examine the relation between teachers activities in the classroom and peacebuilding. REFERENCES Aikman, Sheila, and Elaine Unterhalter Beyond Access: Transforming Policy and Practice for Gender Equality in Education. Oxford, UK: Oxfam Publishing. Anderson, Davie, and Vigids Broch-Due The Poor Are Not Us: Poverty and Pastoralism in Eastern Africa. Oxford, UK, and Athens, OH: James Curry and Ohio University Press. July

74 CHINEN, COOMBES, DE HOOP, CASTRO-ZARZUR, AND ELMESKI Annan, Jeannie, Christopher Blattman, Dyan Mazurana, and Kristopher Carlson Civil War, Reintegration, and Gender in Northern Uganda. Journal of Conflict Resolution 55 (6): Ashmore, Richard, Frances Del Boca, and Scott Bilder Construction and Validation of the Gender Attitude Inventory: A Structured Inventory to Assess Multiple Dimensions of Gender Attitudes. Sex Roles 32 (11-12): Baber, Kristine, and Corinna Tucker The Social Roles Questionnaire: A New Approach to Measuring Attitudes toward Gender. Sex Roles 54 (7-8): Baranyi, Stephen, and Kristiana Powell Fragile States, Gender Equality and Aid Effectiveness: A Review of Donor Perspectives. Ottawa: The North- South Institute. Barker, Gary, and Jennifer Schulte Engaging Men as Allies in Women s Economic Empowerment: Strategies and Recommendations for CARE Country Offices. Washington, DC: CARE Norway by International Center for Research on Women. Barr, Abigail, Frederick Mugisha, Pieter Serneels, and Andrew Zeitlin Information and Collective Action in Community-Based Monitoring of Schools: Field and Lab Experimental Evidence from Uganda. Unpublished draft, Dartmouth College, Hanover, NH. Blattman, Christopher, and Jeannie Annan The Consequences of Child Soldiering. Review of Economics and Statistics 92 (4): Bonder, Gloria Altering Sexual Stereotypes through Teacher Training. In Women and Education in Latin America: Knowledge, Power, and Change, edited by Nelly P. Stromquist. Boulder, CO: Lynne Rienner, Burde, Dana, Ozen Guven, Jo Kelcey, Heddy Lahmann, and Khaled Al-Abbadi What Works to Promote Children s Educational Access, Quality of Learning, and Wellbeing in Crisis-Affected Contexts. Education Rigorous Literature Review, Department for International Development. London: Department for International Development. 74 Journal on Education in Emergencies

75 CAN TEACHER-TRAINING PROGRAMS INFLUENCE GENDER NORMS? Burde, Dana, and Leigh Linden Bringing Education to Afghan Girls: A Randomized Controlled Trial of Village-Based Schools. American Economic Journal: Applied Economics 5 (3): Chinen, Marjorie, Andrea Coombes, Thomas De Hoop, and Mohammed Elmeski Final Report for the Impact Evaluation of the Transformative Potential of Positive Gender Socialization in Education for Peacebuilding in Uganda. Washington, DC: American Institutes for Research. Evaluation-of-the-Transformative-Potential-of-Positive-Gender-Socialization. pdf. Conn, Katherine Identifying Effective Education Interventions in Sub- Saharan Africa: A Meta-Analysis of Rigorous Impact Evaluations. New York: Columbia University Press. De Hoop, Thomas, Luuk van Kempen, Rik Linssen, and Anouka van Eerdewijk Women s Autonomy and Subjective Well-Being: How Gender Norms Shape the Impact of Self-Help Groups in Odisha, India. Feminist Economics 20 (3): Development Research and Training Gender, Conflict, and Peacebuilding Teacher Handbook. Kampala: UNICEF Uganda. Duncan, Greg J The Contributions of Hard Skills and Socio-Emotional Behavior to School Readiness. Unpublished doctoral dissertation, Northwestern University, Chicago. Dworkin, Shari L., Megan S. Dunbar, Suneeta Krishnan, Abigail M. Hatcher, and Sharif Sawires Uncovering Tensions and Capitalizing on Synergies in HIV/AIDS and Antiviolence Programs. American Journal of Public Health 101 (6): Glaser, Barney, and Anslem Strauss The Discovery of Grounded Theory: Strategies for Qualitative Research. Chicago: Alpine. Glick, Peter, and Susan T. Fiske Hostile and Benevolent Sexism Measuring Ambivalent Sexist Attitudes toward Women. Psychology of Women Quarterly 21 (1): July

76 CHINEN, COOMBES, DE HOOP, CASTRO-ZARZUR, AND ELMESKI Instituto da Defesa Nacional Gender Violence in Armed Conflicts (Report no. 11). Lisbon, Portugal: Instituto da Defesa Nacional. Karlan, Dean, Margaret McConnell, Sendhil Mullainathan, and Jonathan Zinman Getting on the Top of Mind: How Reminders Increase Savings (NBER working papers 16205). Cambridge, MA: National Bureau of Economic Research. Kazianga, Harounan, Dan Levy, Leigh Linden, and Matt Sloan The Effects of Girl-Friendly Schools: Evidence from the BRIGHT School Construction Program in Burkina Faso. American Economic Journal: Applied Economics 5 (3): Knutzen, Anna, and Alan Smith Uganda Conflict Analysis (Internal document). New York: UNICEF. McEwan, Patrick J Improving Learning in Primary Schools of Developing Countries: A Meta-Analysis of Randomized Experiments. Review of Educational Research 85 (3): McKenzie, David Beyond Baseline and Follow-Up: The Case for More T in Experiments. Journal of Development Economics 99 (2): Ministry of Gender, Labour, and Social Development, and UNICEF Situation Analysis of Children in Uganda (Report). Kampala: UNICEF Uganda. Mkutu, Kennedy Guns and Governance in the Rift Valley: Pastoralist Conflict and Small Arms, African Issues. Bloomington: Indiana University Press. Mukhopadhyay, Maitrayee, and Franz Wong Revisiting Gender Training: The Making and Remaking of Gender Knowledge. A Global Sourcebook. Oxford, UK: Oxfam Publishing. Ndiwalana, Ali Uganda Case Study: Transformation-Ready. The Strategic Application of Information and Communication Technologies in Africa (Final education sector study). London: ICT Development Associates Ltd. 76 Journal on Education in Emergencies

77 CAN TEACHER-TRAINING PROGRAMS INFLUENCE GENDER NORMS? Opinia, Sylvia, and Friederike Bubenzer Gender Justice and Reconciliation in Northern Uganda (Policy brief no. 4.) Cape Town: Institute for Justice and Reconciliation. Pham, Phuong, and Patrick Vinck Transitioning to Peace: A Population- Based Survey on Attitudes about Social Reconstruction and Justice in Northern Uganda. Berkeley, University of California, Human Rights Center. Pham, Phuong N., Patrick Vinck, and Niamh Gibbons Population-Based Survey on Peace and Education: Uganda. New York: UNICEF, Learning for Peace, Harvard Humanitarian Initiative. Pulerwitz, Julie, and Gary Barker Measuring Attitudes toward Gender Norms among Young Men in Brazil: Development and Psychometric Evaluation of the GEM Scale. Men and Masculinities 10 (3): Puri, Jyotsna, Anastasia Aladysheva, Vegard Iversen, Yoshodhan Ghorpade, and Tilman Brück What Methods May Be Used in Impact Evaluations of Humanitarian Assistance? (Discussion paper no. 8755). Bonn: Institute of Labor Economics. Saminsky, Alina Burying the Pen: Tackling Resistance to Formal Education by the Karamojong in Uganda. Inquiries Journal/Student Pulse 2 (2). Showers, Beverly, and Bruce Joyce The Evolution of Peer Coaching. Educational Leadership 53: Specht, Irma Gender, Disarmament, Demobilization and Reintegration and Violent Masculinities (Report no. 11). Lisboa: IDN Cadernos. Strickland, Richard, and Nata Duvvury Gender Equity and Peacebuilding: From Rhetoric to Reality. Finding the Way (Discussion paper). Washington, DC: International Center for Research on Women. Stromquist, Nelly P The Gender Socialization Process in Schools: A Cross- National Comparison (Background paper prepared for the Education for All Global Monitoring Report 2007). New York: UNESCO. July

78 CHINEN, COOMBES, DE HOOP, CASTRO-ZARZUR, AND ELMESKI Sullivan, Oriel What Do We Learn about Gender by Analyzing Housework Separately from Child Care? Some Considerations from Time Use Evidence. Journal of Family Theory & Review 5 (2): Sustainable Development Goals Goal 5: Achieve Gender Equality and Empower all Women and Girls. wp-content/uploads/2016/08/5_why-it-matters_genderequality_2p.pdf. Vaughan, Jenny, and Tim Stewart Cattle Raiding in Karamoja: A Conflict and Market Assessment. Portland, OR: Mercy Corps. White, Howard Theory-Based Impact Evaluation: Principles and Practice. Journal of Development Effectiveness 1 (3): White, Howard, and Daniel Phillips Addressing Attribution of Cause and Effect in Small n Impact Evaluations: Towards an Integrated Framework (Working paper 15). New Delhi: International Initiative for Impact Evaluation. Winthrop, Rebecca, and Jackie Kirk Learning for a Bright Future: Schooling, Armed Conflict, and Children s Well Being. Comparative Education Review 52 (4): World Bank Gender in Agriculture Sourcebook. Washington, DC: World Bank. 78 Journal on Education in Emergencies

79 THE LIMITS OF REDISTRIBUTIVE SCHOOL FINANCE POLICY IN SOUTH AFRICA Rachel Hatch, Elizabeth Buckner, and Carina Omoeva ABSTRACT Since the end of apartheid, South Africa has embarked on extensive reforms aimed at promoting social cohesion, including progressive educational finance policy (e.g., the no-fee school policy) intended to redress historical inequalities. Because improving equality in and through education is vital to social cohesion, this case study examines whether the no-fee school policy has equalized or is perceived to have equalized school resources and educational opportunities in basic education. Using a mixed-methods approach that draws on household and school survey data and in-depth interviews, we find that the no-fee school policy has reduced the financial burden on black South Africans but that wide gaps in school resources remain. Moreover, we find that the concentration of black students in schools in the poorest areas and of white students in schools in the wealthiest areas rose between 2003 and 2013, and that some black South Africans are dissatisfied with their poor access to elite schools and the superior educational opportunities they offer. Our study argues that South Africa s current school finance policies may be better characterized as pro-poor than redistributive, and points to implications for social cohesion. INTRODUCTION During apartheid, South Africa institutionalized race-based inequalities throughout society, including in its education system. Due to the highly unequal and exclusionary structures and practices of apartheid, strengthening social cohesion in the South African context required addressing past inequities. In the post-apartheid transition to democracy, the government s broader efforts have Received September 2, 2016; revised March 27, 2017, and June 2, 2017; accepted June 13, 2017; electronically published July Journal on Education in Emergencies, Vol. 3, No. 1 Copyright 2017 by the Inter-Agency Network for Education in Emergencies (INEE). All rights reserved. ISSN July

80 HATCH, BUCKNER, AND OMOEVA taken the form of specific policies to provide redress and redistribution. The education system has been a key channel for these efforts, and the government has sought to transform a deeply divided and unequal education system into an equitable one. As such, South Africa makes an excellent case for exploring how educational policy targeting inequality plays into a broader peace and reconciliation process. Specifically, this study centers on how changes to the no-fee school policy instituted after 2010 have shaped the policy s implementation and, ultimately, its contributions to equity. We argue that the policy may be better characterized as pro-poor than redistributive. The no-fee school policy has reduced the financial burden on many black households, which are often in poorer communities, but the ability of schools in wealthier areas to strengthen their budgets through higher school fees and other strategies has meant that inequalities in school resources remain. It appears that, because racial inequalities in South Africa overlap with socioeconomic, linguistic, and geographic divisions, the ability of redistributive educational policies to effect equity continues to be limited. EDUCATION AND SOCIAL COHESION Social cohesion can be understood as the quality of coexistence between the multiple groups that operate within a society (UNICEF 2016, para. 3). Stewart (2014) proposes that social cohesion is a product of three components: (a) low levels of inequality and marginalization; (b) stable, positive social bonds; and (c) an inclusive national identity. Berger-Schmitt (2002) conceptualizes only two dimensions: (a) low levels of inequality; and (b) social capital, which refers to durable, positive interactions and relationships across societies. In conflict and postconflict contexts that have been built on highly unequal and exclusionary policies, redistribution is necessary to effectively lower levels of inequality (Fraser 2005; Novelli 2016; Novelli, Lopes Cardozo, and Smith 2015). Through redistribution, fraught societies equalize resources and opportunities and counteract legacies of oppression and inequity. Inequality plays a central role in these frameworks because improving social cohesion requires addressing the sources of conflict, and mounting evidence points to inequality as a common root cause (Alesina and Perotti 1996; Bartusevičius 2014). Inequality is theorized to be a powerful driver of conflict, particularly where inequalities fall along racial, ethnic, or religious group lines (often termed horizontal inequality). This is because inequality may fuel grievances, which provide a motive, while group 80 Journal on Education in Emergencies

81 THE LIMITS OF REDISTRIBUTIVE SCHOOL FINANCE POLICY IN SOUTH AFRICA dynamics may facilitate mobilization for conflict (Brinkman, Attree, and Hezir 2013; Stewart 2000). It is worth distinguishing between equality, where all groups are treated equally regardless of their differing circumstances, and equity, which recognizes that unequal treatment is often necessary in the pursuit of justice. In this article, we use the words equality or inequality to refer to objective differences in groups access to school or educational funding. In contrast, we use the words equity and inequity to refer to policies of unequal treatment, which either advance social justice or discriminate against certain groups, respectively. While economic, political, social, and cultural inequalities can all spark conflict (Stewart 2008), education warrants critical attention as a force that shapes inequality and violence and equality and social cohesion. First, inequalities in education are themselves problematic and an impediment to social cohesion (Novelli 2016). Second, inequalities in other domains arise through education because of links between education and employment opportunities, social standing, and political participation (Brown 2011; Novelli 2016). Empirical support for this argument is growing, and it now includes recent cross-national time series analyses showing that countries with higher levels of educational inequality across identity groups are more likely to experience conflict (Omoeva and Buckner 2015; Østby 2007, 2008). On the other hand, this means that education systems also have the power to advance equity, not only in education but throughout society. For example, education policies may aim to improve educational outcomes and, consequently, economic and social opportunity for disadvantaged students through policies such as the elimination of school fees, improving the educational infrastructure, or ensuring that schools support the linguistic diversity of their students (Bush and Saltarelli 2000; Smith, McCandless, Paulson, and Wheaton 2011). In education finance, equality indicates equal funding for all students, while pro-equity policies are typically progressive in nature, which includes granting additional government funds to the neediest or historically marginalized populations. It is this potential for education to systematically reduce inequalities that led Novelli (2016) and Novelli, Lopes Cardozo, and Smith (2015) to argue that the redistribution of educational opportunities is one of the primary mechanisms for achieving greater equity. July

82 HATCH, BUCKNER, AND OMOEVA In this study, we attend to the potential of redistributional education policy in South Africa to advance social cohesion by improving equity. We recognize that equity is necessary, but by itself is not a sufficient condition for social cohesion. We acknowledge further that equity-enhancing policies are not only essential to social cohesion but also politically sensitive. Even effective, well-intentioned policies may reignite grievances or violence if they are seen as unfairly privileging some over others, even where new benefits are meant to correct deep disadvantages (Brown 2011; Davies 2010). One example is affirmative action in education, which may increase educational opportunities for disadvantaged groups but also risks accentuating group divisions (Stewart, Brown, and Langer 2007). Such a policy may also be inflammatory when groups that are better off perceive it as unjust, which exemplifies the important point that perceived inequality can be more powerful than actual inequality (Stewart 2008). For this reason, our case study examines the relationship of South Africa s no-fee school policy to substantive shifts in inequality, and to perceptions of how equitable the policy is. REDISTRIBUTIVE EDUCATIONAL POLICIES IN POST-APARTHEID SOUTH AFRICA Apartheid in South Africa was an official policy of racial separation that lasted from 1948 to The Population Registration Act of 1950 classified individuals according to four racial categories white, black, colored, and Indian and mandated divisions across society, including in education, housing, employment, and marriage (Clark and Worger 2004). The 1953 Bantu Education Act strictly segregated schools and differentiated curricular content to suppress the educational opportunity of black South Africans. Public financing for education was also highly unequal. For example, in 1986, subsidies for white students were more than four times higher than those for black students (Vally 1999). Owing in part to community protests and demands, the government of South Africa has significantly reformed education since the fall of apartheid, including systematically dismantling the segregated system (Kallaway, Kruss, Fataar, and Donn 1997; Motala and Pampallis 2001; Sayed, Subrahmanian, Soudien, and Carrim 2007). The right to a free basic public education was enshrined in the South African Constitution (Ahmed and Sayed 2009). The new government abolished what were previously racially separate education departments to create a unified department organized by province (Christie 2006; Fiske and Ladd 2004). Legal segregation was outlawed, and students were permitted to enroll in any school regardless of their race, provided there was space for them. The government also 82 Journal on Education in Emergencies

83 THE LIMITS OF REDISTRIBUTIVE SCHOOL FINANCE POLICY IN SOUTH AFRICA passed a series of sweeping reforms to address social cohesion, expand access, and stem discrimination. These reforms included revised curricular content that eliminated racist language and promoted a unified national identity, and reformed teacher development and deployment and school-based programming to promote social cohesion (Chisholm 2003, 2004; Jansen 1999). A full treatment of South Africa s broader reforms to support social cohesion is outside the scope of this article, as we focus more narrowly on education finance as a key redistributional policy. Given the highly unequal financing within the education system during the apartheid era, an explicit goal of post-apartheid education policies has been to address educational inequality. As Mestry (2014) explains, One of the chief objectives of South Africa s government for the past eighteen years has been to improve the conditions in public schools by diminishing inequalities that exist between schools (852). In this section, we discuss changes in resource allocation to schools, which is one of the primary policy arenas South Africa targeted for redistribution. Table 1 provides an overview of this policy, its evolution, and other key education policies referenced in our analysis. Table 1: Overview of School Finance and Education Policies Discussed in This Study Year Educational Policy Policy Goals 1990 White schools permitted to allow black student enrollment under specific conditions To facilitate partial integration of schools under strict conditions, including the maintenance of a majority white student body and of the white cultural ethos of the school (Vally and Dalamba 1999, 10) Official desegregation of all schools To allow schools to legally diversify; to remove legal barriers to desegregation (Vally and Dalamba 1999, 10) 1994 Teacher Post Provisioning Equalize student-teacher ratios nationwide 1996 South African Schools Act Establish school governing bodies that are allowed to levy compulsory school fees 1998 National Norms and Standards for School Funding (NNSSF) Establish policies guiding school funding. Directed provinces to spend 60 percent of educational budget on poorest 40 percent of schools Fee exemptions Amendment to SASA introduces school fee exemptions based on a means test for low-income households Amended National Norms and Standards for School Funding (ANNSSF) Establish schools in Quintiles 1 and 2 as no-fee schools 2010 Quintile 3 Expansion Expand no-fee school status to all schools in wealth Quintile 3 Early on, the post-apartheid government introduced efforts to promote local governance and redistribute resources in educational settings, including through policies governing school funding and teacher deployment. The 1996 South African Schools Act (SASA) decentralized control of the education system and required July

84 HATCH, BUCKNER, AND OMOEVA all schools to create a democratically elected school governing body (SGB) (Hill 2016). Under SASA, the SGBs were encouraged to supplement official funds with outside funding, such as charging school fees. Nearly all students in South African schools were expected to pay school fees, despite provisions stating that students could not be denied access based on the inability to pay (Hill 2016). Given that the government covered the cost of teacher salaries, school fees were estimated to account for only a small portion of the overall operating budget. Nonetheless, they were thought to contribute to enormous inequities between schools (Motala and Sayed 2012, 20), as wealthy white families were able to pay substantially higher school fees than the historically poor and marginalized groups. In short, although SASA aimed to empower local communities in governance and decision-making, given the deep inequalities in local communities economic resources across South Africa, two major issues arose: first, many poor families simply could not afford school fees and, despite legal protections, were denied access; second, school fees led to significant differences in schools actual resources. In 1998, recognizing that not all parents could afford school fees, the government passed an amendment to SASA that exempted parents from paying fees by introducing waivers. The policy established a means test for fee exemptions based on a family s total income. To compensate schools for the waived fees, the government introduced a per-student allowance for each qualifying student, which did not necessarily reflect the actual school fees. However, only 2 percent of parents actually took advantage of the waivers (Garlick 2013); many others chose noncompliance by refusing to pay fees. In 1998, the education department set new policies for school funding, known as the National Norms and Standards for School Funding (NNSSF). Under NNSSF, all schools were categorized under national wealth quintiles within each province, based on the characteristics of the surrounding community, including unemployment and illiteracy rates; national funding was then allocated on a progressive curve. Motala and Sayed (2012) note the redistributive intent of the policy, stating that it acknowledges that the poor need greater support, but also that the apartheid legacy of poverty remains (23). Nevertheless, the policy still encouraged the collection of fees in schools, which, as it did under SASA, put a significant financial burden on many families and was acknowledged in the policy itself as contributing to resource inequalities (Department of Basic Education 1998). In a large-scale policy revision, the National Norms and Standards for School Funding were amended in 2006 (Ahmed and Sayed 2009). The revised version, the 84 Journal on Education in Emergencies

85 THE LIMITS OF REDISTRIBUTIVE SCHOOL FINANCE POLICY IN SOUTH AFRICA Amended National Norms and Standards for School Funding (ANNSSF), became known as the no-fee school policy because it eliminated fees for schools in the poorest two wealth quintiles, with the government providing higher levels of perstudent funding for non-personnel, non-capital expenditures. Funding allocations for Quintiles 3-5 were lower, and schools in these categories were expected to garner additional funds through fees. That said, the policy only affected nonpersonnel allocation, which was set at 20 percent of total expenditure. In 2010, the ANNSSF policy was reformed to declare Quintile 3 schools no-fee schools. In 2013, the policy was again revised to ensure that all no-fee schools received the same allocation per student per year. In contrast, schools in Quintiles 4 and 5 were still designated as fee-collecting schools, and they received different allocations from the government (Table 2). Under the fee-exemption policy, learners enrolled in fee-paying schools were possibly eligible for a means-tested waiver, and schools received government allocations for each learner who qualified, up to the no-fee school funding level (Department of Basic Education 2015). Table 2: Current Government Allocations to Schools by Quintile School Quintile Per-Student Allocation (ZAR) Quintile 1 1,116 Quintile 2 1,116 Quintile 3 1,116 Quintile Quintile SOURCE: 2015 Department of Basic Education, Amended National Norms and Standards for School Funding. Despite seemingly progressive post-apartheid education policies that have promoted redistribution, scholars have pointed out their limitations in promoting equity. For example, although the stated goal of the SGB policy was to improve local governance and participation in democracy (Ahmed and Sayed 2009), there was a disconnect between the goal and the actual results of the policy (Sayed and Soudien 2005), and between idealist policies and actual experiences (Christie 2006, 379). Scholars have pointed out that school administrators technical capacity challenges made it difficult in some cases to implement post-apartheid era reforms (Christie 2006). Additionally, Ahmed and Sayed (2009) observe that, due to major limitations of census data, it was hard to properly classify schools into quintiles. July

86 HATCH, BUCKNER, AND OMOEVA While significant progress has been made in equalizing access and investments across racial groups (Brook Napier 2005; Chisholm 2003; Christie 1998; Fiske and Ladd 2004; Jansen 2002), apartheid legacies are still strong and opportunities, both in education and more broadly, remain highly unequal (Gilmour and Soudien 2009). This persistent inequality occurs because, when coupled with a macroeconomic environment that does not devote more resources to education, the ability of the no-fee school policy to transform apartheid-era legacies has been limited (Christie 2006; Spreen and Vally 2006). It is ironic that schools in the wealthiest communities (Quintile 5) continue to be able charge school fees, which they use to exclude certain students, including lower class students who live in their vicinity. As a result, private schools and the formerly white elite schools, once termed Model-C schools, continue to offer a higher standard of education, while schools in the townships primarily serve black students and have fewer resources and lower outcomes (Soudien 2004). Given the Model-C schools location in wealthier areas, scholars have argued that class is becoming a more important determinant of access to them than race (Fiske and Ladd 2004; Motala 2006; Soudien 2004, 2007). Our study builds on this rich literature on post-apartheid education reform to further explore the connection of education finance reform to educational equity and social cohesion. Recent studies of the no-fee school policy tend to be literature or policy reviews, rather than empirical studies (see Mestry 2014). Ahmed and Sayed (2009) point to many potential problems with the no-fee school policy, but also argue convincingly that their study must be followed up with empirical data to understand how the policy unfolds practically (214). We bring empirical insights to the study of the no-fee school policy and its impact by drawing on a decade of household survey data and school-level administrative data. Moreover, while numerous empirical studies have examined students school access and outcomes, few studies have interrogated South African citizens perceptions of post-apartheid education reforms more broadly. Gauging citizens perceptions of their educational opportunities and experiences is particularly important to understanding the link between education and social cohesion. The literature suggests that how individuals perceive their opportunities relative to others is at least as important as objective inequalities (Stewart 2008). Moreover, because we have quantitative data from as recently as 2013 and interview data from 2015, we are able to examine attitudes following the 2010 reforms to the quintile classification system, which other studies have not yet examined. 86 Journal on Education in Emergencies

87 THE LIMITS OF REDISTRIBUTIVE SCHOOL FINANCE POLICY IN SOUTH AFRICA DATA AND METHODOLOGY Using a mixed-methods approach, which includes an analysis of nationally representative data sources and qualitative interviews with key stakeholders, this study assesses how the no-fee school policy has been implemented at the school level and to what effect. Specifically, to assess the extent to which the policy has improved equity and, consequently, contributed to social cohesion, we examine the school fees paid by households, the availability of school resources, and educators, parents, and household heads perceptions of the policy s impact. In our quantitative analysis, we use nationally representative general household surveys (GHS), education management information systems data, and Snap Survey of Ordinary Schools data to look at school access, and at the level and perceptions of school resources and household contributions to education. Because these surveys are conducted regularly and are comparable over time (within sources), we are able to examine shifts in key indicators and to compare estimates before and after implementation of the no-fee policy. Our analysis estimates national means using microdata from these data sources (applying survey weighting for the GHS data) and disaggregates by racial group and school quintile wherever possible. To triangulate our quantitative findings and gain more nuanced insight into the impact the no-fee school policy has had on households and schools, we conducted qualitative fieldwork in Limpopo and Western Cape. These provinces were purposively sampled to maximize diversity, which typically is referred to as most different case selection, an approach that is invaluable to understanding heterogeneous settings like South Africa (Seawright and Gerring 2008). The two provinces are among the wealthiest (Western Cape) and poorest (Limpopo) provinces in the country. Limpopo is one of the most racially homogeneous provinces, whereas Western Cape is the only province in which the largest population group (whites) is a minority group at the national level. Table 3 profiles the two provinces. In our interviews, we spoke with key stakeholders who had different perspectives on the education system: education officials, school administrators, teachers, and parents. All interviewees participated voluntarily and gave their informed consent. We worked with Department of Basic Education (DBE) officials in each province to identify participating schools, and visited a total of 19 schools in five districts. 1 School administrators were asked to select teachers and parents for us to speak 1 All interview and focus groups followed IRB guidelines. July

88 HATCH, BUCKNER, AND OMOEVA with. While we aimed to conduct focus groups with a few teachers and parents at each school, the availability differed by school, especially of parents. In total, we interviewed 10 officials, 54 teachers, 20 school administrators, and 24 parents. Our interview and focus group questions asked about individuals and the schools backgrounds, student body composition, and general opinions on the perceived impact of key education policies, especially the no-fee school policy. Following our qualitative fieldwork, we analyzed all interview recordings and identified the key themes that emerged in relation to the no-fee school policy and other policy investments, school resources, and perceptions of equality and social cohesion. As mentioned earlier, this study makes a unique contribution to the literature on the no-fee school policy and its effects by using empirical data especially nationally representative survey data to look at recent progress at the national level. For key indicators of school quality and household contributions to education, it considers both perceptions and actual estimates. In the next section, we discuss our findings on the emerging effects the no-fee school policy has had on education and equity in South Africa and its potential efficacy as a policy that builds social cohesion. Table 3: Overview of Focus Provinces Limpopo Western Cape Population 5,404,868 5,822,734 Racial Breakdown of Population Black/African 96.7% Coloured 0.3% Indian/Asian 0.3% White 2.6% Other 0.2% GDP per capita (USD) 4,259 8,694 Enrollment in Quintile % 39.5 % Schools Black/African 32.8% Coloured 48.8 % Indian/Asian 1.0% White 15.7% Other 1.6% SOURCE: 2011 Census (Statistics South Africa 2012); Statistics South Africa 2011; authors calculations using 2013 Snap Survey of Ordinary Schools data. FINDINGS This section presents our findings on the effectiveness of South Africa s no-fee school policy as a redistribution policy. We begin by addressing the effects the policy has had on school fees and looking at South Africans citizens perceptions of burden, and then discuss its influence on school resources. Finally, we examine issues that have compromised the policy s ability to impact equity and, as a likely result, social cohesion. 88 Journal on Education in Emergencies

89 THE LIMITS OF REDISTRIBUTIVE SCHOOL FINANCE POLICY IN SOUTH AFRICA A Pro-Poor Policy: Alleviating Financial Burden Household Contributions to Education To begin, we examine the effect the no-fee school policy has had on parental contributions to education. Figure 1 shows the percentage of South Africans, by racial group, who reported paying school fees over time, and Figure 2 shows the percentage of individuals who stated that the fees are too high. Prior to the implementation of the no-fee school policy, most South Africans reported paying fees in addition to buying textbooks and uniforms (although our interviews suggest that fee evasion was widespread). The fees posed a significant burden on many families. According to the DBE, in 2003, 56 percent of households in the poorest quintile stated that their children dropped out of school because of the cost of school fees (2009a). By 2011, after no-fee school status was extended to include Quintile 3 schools, the percentage of black South Africans paying fees went down to only 32.4 percent, while more than 95 percent of white South Africans continued to pay school fees. Figure 1: Percentage of Primary and Secondary School Students Whose Households Pay School Fees; Figure 2: Percentage of Primary and Secondary School Students Who Experience School Fees That Are Too High SOURCE: Authors calculations using General Household Survey data. In terms of the strain placed on families, black South Africans, compared to other groups, perceived school fees to be the most burdensome in 2003, with 19 percent saying that school fees were too high. In contrast, only 8 percent July

90 HATCH, BUCKNER, AND OMOEVA of white South Africans said the same. By 2011, these perceptions seemed to have been reversed, with only 3.9 percent of black South Africans arguing that school fees were too high, compared to 15 percent of white South Africans. In qualitative interviews we conducted in Quintile 1-3 schools in both Western Cape and Limpopo provinces, teachers and administrators overwhelmingly agreed that the elimination of school fees was widely lauded by communities, as it reduced the financial burden on parents. This shift in attitudes reflects the actual amount of fees paid. Table 4 shows how much families spent on school fees at the primary and secondary level by year (adjusted for inflation). Both white and colored families were paying higher fees in 2011 than in 2003, with white families seeing the steepest increase. It is clear that white South Africans not only continued to pay fees, the average amount of their fees increased over time. For white and colored families, this may indicate a shift toward enrolling their children in schools at the upper end of the spectrum or in independent schools. Table 4: Average Primary and Secondary School Fees (ZAR) by Race and Year (adjusted for inflation) Race African/Black Coloured White 5, , SOURCE: Authors calculations using General Household Survey data. These changes in perception may have important implications for social cohesion. Black South Africans clearly perceive themselves as better off in 2011 because of the no-fee school policy. On the other hand, white South Africans, who are less likely to attend no-fee schools, have not only seen no benefit but have seen to their dissatisfaction the average amount they pay in school fees rise. School Resources and Environment One of the most important benefits of the no-fee school policy is that it has allowed schools to have a stable and predictable baseline of resources to pay for non-salary recurrent costs, including teaching and learning materials. After 1994, it was illegal to exclude students based on their ability to pay, and the high rates of noncompliance meant that schools operated with severely limited budgets 90 Journal on Education in Emergencies

91 THE LIMITS OF REDISTRIBUTIVE SCHOOL FINANCE POLICY IN SOUTH AFRICA and had limited resources for upkeep, maintenance, and learning materials. For example, teachers we interviewed explained that, before their school became a no-fee school, parents did not pay or paid only part of the school fees, which reduced the school s income. These findings align with a 2009 study conducted by the DBE (2009b), which found that 71 percent of the surveyed no-fee schools indicated that they are able to provide better services with their school allocations as compared to when they were collecting fees (6). Thus, the no-fee school policy can be linked directly to improved school resources in lower income communities. These interview findings are supported by nationally representative data on household heads perceptions of their children s school environment. Figure 3 presents the percentage of South Africans who stated that their school lacks books; in 2003, almost 25 percent of black South African students did not have textbooks, compared to only 1.8 percent of white students. By 2011, this percentage had fallen significantly to only 6.3 percent of black South Africans and had also fallen slightly for colored South Africans. In contrast, the percentage of white South African students who did not have textbooks actually increased modestly to 3.7 percent, although they were the least likely to state that they experienced a lack of books. Figure 4 shows similarly that South Africans perceptions of their school facilities also improved, particularly among black South Africans: the percentage of black South Africans who felt that their school had bad facilities dropped from 12.3 percent in 2003 to 4.4 percent in Figure 3: Percentage of Primary and Secondary School Students Experiencing a Lack of Textbooks; Figure 4: Percentage of Primary and Secondary School Students Experiencing School Facilities in Bad Condition SOURCE: Authors calculations using General Household Survey data July

92 HATCH, BUCKNER, AND OMOEVA Teacher Availability We also looked at pupil-teacher ratios in public schools by wealth quintile at two points in time, 2006 and Table 5 shows that pupil-teacher ratios have declined slightly for public schools in all wealth quintiles, and that pupil-teacher ratios in no-fee schools are not substantially different from those in Quintile 4 fee-paying schools. However, schools in Quintile 5 still have lower pupil-teacher ratios than those in other quintiles. In fact, in accordance with SASA, fee-paying schools can hire additional teachers with the funds they generate from fees and outside fundraising. For example, we visited a Quintile 5 school where the SGB paid for 16 full-time teachers. Having the additional teachers helped to keep class size small and manageable, which our interviewees said created strong disparities in the quality of teaching students received. Table 5: Pupil-Teacher Ratios by School Quintile and Year Quintile 1 Quintile 2 Quintile 3 Quintile 4 Quintile SOURCE: Authors calculations using Snap Survey of Ordinary Schools data. In short, it is clear that the no-fee policy has helped provide a minimum resource base for all schools and reduced the burden of paying fees, improved the school environment, and increased the availability of teachers over the past decade. While GHS and Snap Survey of Ordinal Schools data do not allow us to attribute changes to a particular policy, they do demonstrate that changes register at the national level. Responses from the interviews provide more policy insight and suggest that the ANNSSF has been a key factor in easing the burden of school fees, especially in black households, and has played a role in providing schools with a baseline of resources. We argue, therefore, that the no-fee school policy can be considered reasonably effective as a pro-poor policy, in that it has had a differentially large and positive impact on the poorest and historically marginalized populations. However, its potency as an equity-building policy (and thus a policy that advances social cohesion) is less clear and the topic we turn to next. 2 Although we would like to examine the change over a longer period, we do not have teacher data prior to Journal on Education in Emergencies

93 THE LIMITS OF REDISTRIBUTIVE SCHOOL FINANCE POLICY IN SOUTH AFRICA Little Actual Redistribution Despite some successes in alleviating the financial burden of fees for black families and improving educational resources in poorer areas, our findings including the finding above that better pupil-teacher ratios have been maintained in Quintile 5 schools indicate that the no-fee school policy has not had a substantive impact on equity. In this section, we discuss the fact that, despite its progressive funding allocation, the no-fee school policy has been unable so far to address the systemic inequalities affecting education in South Africa, including (a) persistent gaps in school resources, owing in part to the schools different opportunities to raise funds through fees and other means; and (b) de facto housing segregation. Persistent Gaps in School Resources Although the no-fee policy has redistributive goals, considerable and predictable resource gaps remain across schools. These gaps stem in part from schools different ability to raise funds through fees in wealthier and poorer communities, which simultaneously allows schools to exclude students who cannot pay fees on the grounds of class, language, or race, and thus exacerbates inequalities in resources. Table 4 shows that white families pay more now than in the past for their children s education, and that the amount they pay is roughly six times what the government provides per pupil in no-fee schools. In the aggregate, this likely means that the absolute difference in school resources between schools with more white students and those serving primarily black students is larger now than before the no-fee school policy was implemented. It is important to point out that this is not due to government policy but to the fact that spending by white (and, for the most part, wealthy) parents has outpaced government investment. While we caution that information on school fees provides only an approximate view of school budgets and that other factors play into school finances, the widening gap in effective school fees paid between 2003 and 2011 clearly suggests that the no-fee school policy has most likely not been able to equalize resources across schools. In addition to widening the gap in family contributions, unintended consequences stemming from issues with the school quintile classification system mean that there are large inequalities in school resources even within Quintile 4 and 5 schools. Prior research has been critical of the poor quality data on quintile classification and its implications for achieving equity (Mestry 2014). Our study found similarly that some fee-charging schools technically classified as Quintile 4 or 5 serve many low-income students. This is because, due to data constraints, the ANNSSF classifies schools into quintiles based on characteristics of the community July

94 HATCH, BUCKNER, AND OMOEVA surrounding the school, rather than on the characteristics of the students who attend the school. As a result, it is not uncommon for fee-paying schools to serve very poor students who happen to live in wealthier catchment areas. In practice, this means that some students who are the intended beneficiaries of ANNSSF do not benefit from it and, as we discuss later, that school budgets suffer when these students cannot or do not pay school fees. Indeed, one of the most surprising findings from our fieldwork was that some Quintile 4 and 5 schools struggle to provide basic teaching and learning resources because they cannot collect fees from all students. For example, one teacher from a school with fees explained that only 62 percent of the expected fees were collected, and another observed that parents have fallen into a culture of nonpayment. When a school is classified as Quintile 4 or 5, the school receives less than half the amount from the government as it would if it were a no-fee school. For schools serving lower income students, this results in their having a smaller operating budget than if they had been designated no-fee schools. A DBE (2009b) article based on surveys with school principals recognized this problem, explaining that most schools in urban areas servicing poor communities lose a large portion of their school income due to fee exemptions granted to poor parents (9). Others have observed this challenge and attributed it to the inaccurate formulation of quintile designations for some schools (Ahmed and Sayed 2009), which our more recent empirical findings affirm to be an ongoing difficulty even with the amendments to the no-fee school policy. For example, the principal of a Quintile 5 secondary school in Western Cape explained that, in his school, it s not strange for the school to have no telephone connection, because we don t have money to pay the telephone [bill]. In a Quintile 4 school, teachers said that many students did not have textbooks because they could not afford to buy them, and the government did not provide them as it did for no-fee schools. As a result, although the school was a Quintile 4, teachers often had to resort to using photocopies for lessons, rather than textbooks. Thus, some fee-paying schools inability to collect fees in full has had a severe impact on their ability to meet even basic expenditures for educational resources. This challenge is exacerbated by the fact that schools also have very different resource legacies for example, better instructional spaces, sports fields, and extracurricular facilities and abilities to mobilize additional financial resources. For example, former Model-C schools, which were previously white-only public schools, have superior resource legacies. They tend to be located in wealthy residential areas with a more affluent student base, and thus are better able to 94 Journal on Education in Emergencies

95 THE LIMITS OF REDISTRIBUTIVE SCHOOL FINANCE POLICY IN SOUTH AFRICA garner additional financial support from alumni communities and other local fundraising sources than schools in poorer communities. Moreover, because school-fee policies are set by the SGB, these schools can use school fees or selective scholarships to enroll a highly selective student body. De Facto Housing and Community Segregation The ability of the no-fee school policy to influence equity in education is further constrained by de facto housing and community segregation and their impact on school access. Homogeneity in communities stems in part from apartheid policies that created racially and ethnically distinct township communities, and in part from newly created government housing settlements for low-income families in some areas. Interview respondents explained that this lack of diversity within communities is reflected in the schools that is, that the tendency for racial or class groups to be spatially concentrated undermines the potential for greater integration in the schools and, consequently, reduces the chances that a larger share of poorer students will be able to access elite schools. In theory, poorer students can access fee-paying schools. According to SASA, students can apply to any school outside their immediate geographic area and, since 1998, the government has given a per-student allowance to fee-paying schools that enroll students who cannot pay the fees. While the central aim of the no-fee school policy is redistribution rather than integration, the overlap of race and class in South Africa means that black students are more likely to qualify for exemptions and, as a result, that the policy has the potential to promote racial integration in fee-paying schools. However, the level of exemption is sometimes lower than the actual amount of the fees collected by a school, which is a disincentive for fee-paying schools to admit students who qualify for exemptions. Coupled with the fact that schools rarely have space for all who apply, the result is that very few students who qualify for school fee exemptions are actually granted admission. In an example offered by one of our interviewees, students in a township area travelled more than seven kilometers to get to a no-fee school, even though the elite school in the area was closer to the informal settlements they lived in. However, because the elite school said it had no space for additional students, students were not able to get exemptions to attend there. One interviewee strongly believed that this technical issue of exemptions was a façade used to allow elite public schools to select their desired student bodies while masking ongoing racial and class-based discrimination. July

96 HATCH, BUCKNER, AND OMOEVA To examine quantitatively how school composition has changed, possibly facilitated by the no-fee school policy, we analyzed Education Management Information Systems data on the racial breakdown of enrollment by school wealth quintile. 3 Results are presented in Figure 5, which shows that the distribution of students has not changed significantly over the past decade. However, there was a higher concentration of black and colored students in the lowest three quintiles in 2013 than before the ANNSSF was passed. Black students are proportionally more likely to attend Quintile 1 and 2 schools than Quintile 3-5 schools, and colored students are more likely to attend Quintile 3 and 4 schools than they were before the no-fee school policy. In contrast, the percentage of white students attending Quintile 5 schools, already high in 2003 at nearly 82 percent, reached 87 percent by Moreover, almost no white students were attending schools in Quintiles 1-3 by Figure 5: Distribution of Primary and Secondary School Enrollment by Race, Year, and School Wealth Quintile SOURCE: Authors calculations using General Household Survey data. Using these data in combination with self-reports on the level of fees students paid (Table 4), it appears that, in the wake of the no-fee school policy, schools in Quintiles 4 and 5 may have raised fees, which pushed less wealthy colored students into Quintile 3 schools and ensured that only a very small percentage of black students could afford to enroll in the fee-paying schools. Even if the fees were raised purely in the name of higher quality, one outcome has been to maintain the disparity between students of different racial groups. 3 Data provided by EMIS Unit/DBE through personal communication, March 26, Journal on Education in Emergencies